diff --git a/CHANGELOG.d/analysis-run-collection-cli.md b/CHANGELOG.d/analysis-run-collection-cli.md
new file mode 100644
index 000000000..abd6d8431
--- /dev/null
+++ b/CHANGELOG.d/analysis-run-collection-cli.md
@@ -0,0 +1 @@
+- `tepp_api` loopback `tepp-analysis-runs list` enumerates metric-free accepted, running, cancelled, and terminal runs (ADR 0032). Collection CLI stdout refuses RMSE/bias/coverage/SE-gate/scientific-acceptance keys. Not GET-by-id, not scientific-acceptance CLI, not persistence.
diff --git a/CHANGELOG.md b/CHANGELOG.md
index b55a4f9c4..3df3d7446 100644
--- a/CHANGELOG.md
+++ b/CHANGELOG.md
@@ -38,6 +38,8 @@ All notable changes to TEPP are documented here. The format follows Keep a Chang
 
 ## [Unreleased]
 
+- `tepp_api` publishes `tepp-analysis-runs list` as the loopback client of `GET /v1/analysis-runs` (ADR 0032). Operators enumerate accepted, running, cancelled, and terminal runs as metric-free collection rows without writing raw HTTP. CLI stdout refuses RMSE/bias/coverage/SE-gate/scientific-acceptance/`terminal_result` keys. Not GET-by-id, not the scientific-acceptance CLI, and not an ADR 0014 claim.
+
 - `tepp_api` serves `GET /v1/analysis-runs` on the shared loopback listener (ADR 0031). Operators enumerate accepted, running, cancelled, and terminal runs as metric-free collection rows. Collection bodies refuse RMSE/bias/coverage/SE-gate/scientific-acceptance/`terminal_result` keys. GET-by-id and running/terminal POST remain later GAP-003A slices; this is not an ADR 0014 claim.
 
 - `tepp_api` serves `POST /v1/analysis-runs/{run_id}/cancel` on the shared loopback listener (ADR 0029). Accepted and running runs become metric-free `cancelled` status. Succeeded, failed, and unknown runs cannot be cancelled. Cancel bodies refuse RMSE/bias/coverage/SE-gate/scientific-acceptance keys. GET status and running/terminal POST remain later GAP-003A slices; this is not an ADR 0014 claim.
diff --git a/DOCUMENTATION.md b/DOCUMENTATION.md
index b29d53702..ca980638c 100644
--- a/DOCUMENTATION.md
+++ b/DOCUMENTATION.md
@@ -15,6 +15,7 @@ TEPP's approved PRD v0.4 and implementation plan are the primary product baselin
 | Orchestrator live HTTP doctoring | [`docs/research/orchestrator-live-http.md`](docs/research/orchestrator-live-http.md) |
 | Analysis-run cancel HTTP doctoring | [`docs/research/analysis-run-cancel-http.md`](docs/research/analysis-run-cancel-http.md) |
 | Analysis-run collection HTTP doctoring | [`docs/research/analysis-run-collection-http.md`](docs/research/analysis-run-collection-http.md) |
+| Analysis-run collection CLI doctoring | [`docs/research/analysis-run-collection-cli.md`](docs/research/analysis-run-collection-cli.md) |
 | UML/runtime/scientific flows | [`docs/UML.md`](docs/UML.md) |
 | Logical/physical ERD | [`docs/ERD.md`](docs/ERD.md) |
 | Security policy | [`SECURITY.md`](SECURITY.md) |
diff --git a/crates/tepp_api/Cargo.toml b/crates/tepp_api/Cargo.toml
index 47ad7c433..b3c3d43f4 100644
--- a/crates/tepp_api/Cargo.toml
+++ b/crates/tepp_api/Cargo.toml
@@ -29,5 +29,11 @@ path = "src/bin/tepp_loopback.rs"
 test = false
 bench = false
 
+[[bin]]
+name = "tepp-analysis-runs"
+path = "src/bin/tepp_analysis_runs.rs"
+test = false
+bench = false
+
 [lints]
 workspace = true
diff --git a/crates/tepp_api/src/analysis_run_collection_cli.rs b/crates/tepp_api/src/analysis_run_collection_cli.rs
new file mode 100644
index 000000000..4d02bd1ac
--- /dev/null
+++ b/crates/tepp_api/src/analysis_run_collection_cli.rs
@@ -0,0 +1,890 @@
+//! Operator loopback CLI for analysis-run collection GET.
+//!
+//! GAP-003A eighth slice: operators run `tepp-analysis-runs list` to enumerate
+//! accepted, running, cancelled, and terminal runs without writing raw HTTP or
+//! guessing run identities. Stdout stays metric-free. `tepp.scientific_acceptance.v1`
+//! never appears. This module does not duplicate GET-by-id (#359), lifecycle
+//! POST (#360), cancel HTTP (#361), scientific-acceptance CLI (#362), or the
+//! collection GET listener (#368). Persistence remains GAP-003B.
+
+use std::io::{Read, Write};
+use std::net::{SocketAddr, TcpStream};
+
+use crate::analysis_run_collection_http::{
+    parse_collection_page_cursor, parse_collection_page_limit, refuse_metrics_on_collection_payload,
+};
+use crate::lineageweave_http::consumer_is_supported;
+use crate::live_http::map_io_error;
+use crate::naruon_http::{NARUON_ANALYSIS_RUN_PATH, header_is_credential};
+use crate::wire::require_nonempty;
+use crate::{
+    AnalysisRunCollection, AnalysisRunLiveService, ApiError, NARUON_LIVE_IO_TIMEOUT,
+    NaruonLiveResponse,
+};
+
+const SCIENTIFIC_ACCEPTANCE_SCHEMA: &str = "tepp.scientific_acceptance.v1";
+
+/// Supported operator verbs for the loopback collection CLI.
+#[derive(Clone, Copy, Debug, Eq, PartialEq)]
+pub enum AnalysisRunCollectionCliVerb {
+    /// `GET /v1/analysis-runs`.
+    List,
+}
+
+impl AnalysisRunCollectionCliVerb {
+    /// Parse one exact lowercase verb token.
+    ///
+    /// # Errors
+    ///
+    /// Returns [`ApiError::InvalidWirePayload`] for an unknown token.
+    pub fn parse(token: &str) -> Result<Self, ApiError> {
+        match token {
+            "list" => Ok(Self::List),
+            _ => Err(ApiError::InvalidWirePayload),
+        }
+    }
+
+    /// Return the canonical lowercase verb token.
+    #[must_use]
+    pub const fn as_str(self) -> &'static str {
+        match self {
+            Self::List => "list",
+        }
+    }
+}
+
+/// One operator CLI invocation against a loopback collection GET listener.
+#[derive(Clone, Debug, Eq, PartialEq)]
+pub struct AnalysisRunCollectionCliInvocation {
+    /// CLI verb to execute.
+    pub verb: AnalysisRunCollectionCliVerb,
+    /// Loopback `host:port` of `tepp-loopback`.
+    pub host: String,
+    /// Published modular consumer (`naruon` or `lineageweave`).
+    pub consumer: String,
+    /// Optional exclusive page cursor (`tepp-page-cursor`).
+    pub page_cursor: Option<String>,
+    /// Optional page limit (`tepp-page-limit`).
+    pub page_limit: Option<String>,
+    /// JSON body. Collection GET requires empty.
+    pub body: String,
+}
+
+impl AnalysisRunCollectionCliInvocation {
+    /// Parse argv plus stdin body into a validated loopback collection invocation.
+    ///
+    /// # Errors
+    ///
+    /// Returns a fail-closed error for unknown verbs, missing required flags, a
+    /// non-loopback host, an unpublished consumer, credential-shaped flags,
+    /// hostile pagination, or a nonempty body.
+    pub fn from_args<I, S>(args: I, body: impl Into<String>) -> Result<Self, ApiError>
+    where
+        I: IntoIterator<Item = S>,
+        S: AsRef<str>,
+    {
+        let tokens: Vec<String> = args
+            .into_iter()
+            .map(|token| token.as_ref().to_owned())
+            .collect();
+        let (verb_token, rest) = tokens.split_first().ok_or(ApiError::InvalidWirePayload)?;
+        let verb = AnalysisRunCollectionCliVerb::parse(verb_token)?;
+        let flags = parse_flags(rest)?;
+        assemble_invocation(verb, flags, body.into())
+    }
+
+    /// Reject a non-loopback host, unpublished consumer, or hostile page flags.
+    ///
+    /// # Errors
+    ///
+    /// Returns [`ApiError::AuthorizationDenied`] for a non-loopback host and
+    /// [`ApiError::InvalidWirePayload`] or [`ApiError::LimitExceeded`] for
+    /// empty, unpublished, or out-of-bounds fields.
+    pub fn validate(&self) -> Result<(), ApiError> {
+        require_loopback_host(&self.host)?;
+        require_nonempty(&self.consumer)?;
+        if !consumer_is_supported(&self.consumer) {
+            return Err(ApiError::InvalidWirePayload);
+        }
+        if !self.body.is_empty() {
+            return Err(ApiError::InvalidWirePayload);
+        }
+        refuse_metrics_on_collection_payload(&self.body)?;
+        parse_collection_page_limit(self.page_limit.as_deref())?;
+        parse_collection_page_cursor(self.page_cursor.as_deref())?;
+        Ok(())
+    }
+}
+
+struct ParsedFlags {
+    host: Option<String>,
+    consumer: Option<String>,
+    page_cursor: Option<String>,
+    page_limit: Option<String>,
+}
+
+fn parse_flags(rest: &[String]) -> Result<ParsedFlags, ApiError> {
+    let mut flags = ParsedFlags {
+        host: None,
+        consumer: None,
+        page_cursor: None,
+        page_limit: None,
+    };
+    let mut index = 0;
+    while index < rest.len() {
+        let flag = rest[index].as_str();
+        if !flag.starts_with("--") {
+            return Err(ApiError::InvalidWirePayload);
+        }
+        let name = &flag[2..];
+        if header_is_credential(name) {
+            return Err(ApiError::AuthorizationDenied);
+        }
+        let slot = match name {
+            "host" => &mut flags.host,
+            "consumer" => &mut flags.consumer,
+            "page-cursor" => &mut flags.page_cursor,
+            "page-limit" => &mut flags.page_limit,
+            _ => return Err(ApiError::InvalidWirePayload),
+        };
+        if slot.is_some() || index + 1 >= rest.len() {
+            return Err(ApiError::InvalidWirePayload);
+        }
+        let value = rest[index + 1].as_str();
+        require_nonempty(value)?;
+        *slot = Some(value.to_owned());
+        index += 2;
+    }
+    Ok(flags)
+}
+
+fn assemble_invocation(
+    verb: AnalysisRunCollectionCliVerb,
+    flags: ParsedFlags,
+    body: String,
+) -> Result<AnalysisRunCollectionCliInvocation, ApiError> {
+    let invocation = AnalysisRunCollectionCliInvocation {
+        verb,
+        host: flags.host.ok_or(ApiError::InvalidWirePayload)?,
+        consumer: flags
+            .consumer
+            .unwrap_or_else(|| crate::NARUON_CONSUMER_CODE.to_owned()),
+        page_cursor: flags.page_cursor,
+        page_limit: flags.page_limit,
+        body,
+    };
+    invocation.validate()?;
+    Ok(invocation)
+}
+
+fn require_loopback_host(host: &str) -> Result<SocketAddr, ApiError> {
+    let addr: SocketAddr = host.parse().map_err(|_| ApiError::InvalidWirePayload)?;
+    if addr.ip().is_loopback() {
+        Ok(addr)
+    } else {
+        Err(ApiError::AuthorizationDenied)
+    }
+}
+
+/// Compose one HTTP/1.1 collection GET for a validated CLI invocation.
+///
+/// # Errors
+///
+/// Returns the same fail-closed errors as
+/// [`AnalysisRunCollectionCliInvocation::validate`].
+pub fn compose_analysis_run_collection_cli_http(
+    invocation: &AnalysisRunCollectionCliInvocation,
+) -> Result<String, ApiError> {
+    invocation.validate()?;
+    let mut request = format!(
+        "GET {NARUON_ANALYSIS_RUN_PATH} HTTP/1.1\r\nHost: {}\r\ncontent-type: application/json\r\ntepp-consumer: {}\r\ntepp-contract-version: 1\r\n",
+        invocation.host, invocation.consumer
+    );
+    if let Some(cursor) = &invocation.page_cursor {
+        request.push_str("tepp-page-cursor: ");
+        request.push_str(cursor);
+        request.push_str("\r\n");
+    }
+    if let Some(limit) = &invocation.page_limit {
+        request.push_str("tepp-page-limit: ");
+        request.push_str(limit);
+        request.push_str("\r\n");
+    }
+    request.push_str("content-length: 0\r\n\r\n");
+    Ok(request)
+}
+
+/// Dispatch one collection CLI invocation against an in-process loopback service.
+///
+/// # Errors
+///
+/// Returns fail-closed validation errors before the HTTP handler runs.
+pub fn dispatch_analysis_run_collection_cli(
+    service: &mut AnalysisRunLiveService,
+    invocation: &AnalysisRunCollectionCliInvocation,
+) -> Result<NaruonLiveResponse, ApiError> {
+    let request = compose_analysis_run_collection_cli_http(invocation)?;
+    Ok(service.handle_http_request(&request))
+}
+
+/// Execute one collection CLI invocation over loopback TCP against `tepp-loopback`.
+///
+/// # Errors
+///
+/// Returns fail-closed validation, transport, or response-framing errors.
+pub fn execute_analysis_run_collection_cli(
+    invocation: &AnalysisRunCollectionCliInvocation,
+) -> Result<NaruonLiveResponse, ApiError> {
+    let addr = require_loopback_host(&invocation.host)?;
+    let request = compose_analysis_run_collection_cli_http(invocation)?;
+    let mut stream = TcpStream::connect(addr).map_err(|error| map_io_error(&error))?;
+    stream
+        .set_read_timeout(Some(NARUON_LIVE_IO_TIMEOUT))
+        .map_err(|error| map_io_error(&error))?;
+    stream
+        .set_write_timeout(Some(NARUON_LIVE_IO_TIMEOUT))
+        .map_err(|error| map_io_error(&error))?;
+    stream
+        .write_all(request.as_bytes())
+        .map_err(|error| map_io_error(&error))?;
+    stream.flush().map_err(|error| map_io_error(&error))?;
+    let mut bytes = Vec::new();
+    stream
+        .read_to_end(&mut bytes)
+        .map_err(|error| map_io_error(&error))?;
+    parse_http_response(&bytes)
+}
+
+/// Filter CLI stdout so collection pages never print scientific acceptance.
+///
+/// # Errors
+///
+/// Returns [`ApiError::InvalidWirePayload`] when a receipt carries metric keys
+/// or `tepp.scientific_acceptance.v1`.
+pub fn render_analysis_run_collection_cli_stdout(
+    invocation: &AnalysisRunCollectionCliInvocation,
+    response: &NaruonLiveResponse,
+) -> Result<String, ApiError> {
+    invocation.validate()?;
+    if response.body.is_empty() {
+        return Err(ApiError::InvalidWirePayload);
+    }
+    refuse_scientific_acceptance_schema(&response.body)?;
+    if !(200..300).contains(&response.status_code) {
+        refuse_metrics_on_collection_payload(&response.body)?;
+        return Ok(response.body.clone());
+    }
+    let collection = AnalysisRunCollection::from_json(&response.body)?;
+    collection.to_json()
+}
+
+fn refuse_scientific_acceptance_schema(body: &str) -> Result<(), ApiError> {
+    if body.contains(SCIENTIFIC_ACCEPTANCE_SCHEMA) {
+        Err(ApiError::InvalidWirePayload)
+    } else {
+        Ok(())
+    }
+}
+
+fn parse_http_response(bytes: &[u8]) -> Result<NaruonLiveResponse, ApiError> {
+    let text = std::str::from_utf8(bytes).map_err(|_| ApiError::InvalidWirePayload)?;
+    let (header_block, body) = text
+        .split_once("\r\n\r\n")
+        .ok_or(ApiError::InvalidWirePayload)?;
+    let mut lines = header_block.split("\r\n");
+    let status_line = lines.next().ok_or(ApiError::InvalidWirePayload)?;
+    let mut parts = status_line.split(' ');
+    if parts.next() != Some("HTTP/1.1") {
+        return Err(ApiError::InvalidWirePayload);
+    }
+    let code = parts
+        .next()
+        .ok_or(ApiError::InvalidWirePayload)?
+        .parse::<u16>()
+        .map_err(|_| ApiError::InvalidWirePayload)?;
+    let reason_phrase = static_reason(code)?;
+    let mut content_length = None;
+    for line in lines {
+        let (name, value) = line.split_once(':').ok_or(ApiError::InvalidWirePayload)?;
+        if name.eq_ignore_ascii_case("content-length") {
+            if content_length.is_some() {
+                return Err(ApiError::InvalidWirePayload);
+            }
+            content_length = Some(
+                value
+                    .trim()
+                    .parse::<usize>()
+                    .map_err(|_| ApiError::InvalidWirePayload)?,
+            );
+        }
+    }
+    let declared = content_length.ok_or(ApiError::InvalidWirePayload)?;
+    if declared != body.len() {
+        return Err(ApiError::InvalidWirePayload);
+    }
+    Ok(NaruonLiveResponse {
+        status_code: code,
+        reason_phrase,
+        body: body.to_owned(),
+    })
+}
+
+fn static_reason(code: u16) -> Result<&'static str, ApiError> {
+    match code {
+        200 => Ok("OK"),
+        202 => Ok("Accepted"),
+        400 => Ok("Bad Request"),
+        403 => Ok("Forbidden"),
+        413 => Ok("Payload Too Large"),
+        422 => Ok("Unprocessable Entity"),
+        _ => Err(ApiError::InvalidWirePayload),
+    }
+}
+
+/// Read stdin leftover bytes on a non-terminal; collection GET refuses a body.
+///
+/// # Errors
+///
+/// Returns [`ApiError::InvalidWirePayload`] when stdin cannot be read.
+pub fn read_analysis_run_collection_cli_stdin(
+    stdin_is_terminal: bool,
+    mut stdin: impl Read,
+) -> Result<String, ApiError> {
+    if stdin_is_terminal {
+        Ok(String::new())
+    } else {
+        let mut body = String::new();
+        stdin
+            .read_to_string(&mut body)
+            .map_err(|_| ApiError::InvalidWirePayload)?;
+        Ok(body)
+    }
+}
+
+#[cfg(test)]
+#[allow(clippy::too_many_lines)]
+mod tests {
+    use super::{
+        AnalysisRunCollectionCliInvocation, AnalysisRunCollectionCliVerb,
+        SCIENTIFIC_ACCEPTANCE_SCHEMA, compose_analysis_run_collection_cli_http,
+        dispatch_analysis_run_collection_cli, execute_analysis_run_collection_cli,
+        parse_http_response, read_analysis_run_collection_cli_stdin,
+        render_analysis_run_collection_cli_stdout, static_reason,
+    };
+    use crate::{
+        ANALYSIS_RUN_COLLECTION_MAX_LIMIT, ANALYSIS_RUN_CONTRACT_VERSION, AnalysisRunAccepted,
+        AnalysisRunCollection, AnalysisRunLiveService, AnalysisRunRequest, ApiError,
+        LINEAGEWEAVE_CONSUMER_CODE, NARUON_ANALYSIS_RUN_PATH, NARUON_CONSUMER_CODE,
+        NaruonLiveResponse,
+    };
+
+    fn request(idempotency_key: &str) -> AnalysisRunRequest {
+        AnalysisRunRequest {
+            contract_version: ANALYSIS_RUN_CONTRACT_VERSION,
+            idempotency_key: idempotency_key.into(),
+            tenant_workspace_id: "cli-collection-tenant".into(),
+            snapshot_id: "cli-collection-snapshot".into(),
+            knowledge_cutoff: "2026-08-01T00:00:00Z".into(),
+            model_contract_version: "tepp-analysis-run-v1".into(),
+            output_profile: "calibrated_event_measurement".into(),
+        }
+    }
+
+    fn create_http(run: &AnalysisRunRequest, consumer: &str, host: &str) -> String {
+        let body = run.to_json().expect("json");
+        format!(
+            "POST {NARUON_ANALYSIS_RUN_PATH} HTTP/1.1\r\nHost: {host}\r\ncontent-type: application/json\r\ntepp-consumer: {consumer}\r\ntepp-contract-version: 1\r\nidempotency-key: {}\r\ncontent-length: {}\r\n\r\n{body}",
+            run.idempotency_key,
+            body.len()
+        )
+    }
+
+    fn list_invocation() -> AnalysisRunCollectionCliInvocation {
+        AnalysisRunCollectionCliInvocation::from_args(["list", "--host", "127.0.0.1:18081"], "")
+            .expect("list")
+    }
+
+    #[test]
+    fn verbs_parse_and_reject_unknown_tokens() {
+        assert_eq!(
+            AnalysisRunCollectionCliVerb::parse("list").expect("list"),
+            AnalysisRunCollectionCliVerb::List
+        );
+        assert_eq!(AnalysisRunCollectionCliVerb::List.as_str(), "list");
+        assert_eq!(
+            AnalysisRunCollectionCliVerb::parse("LIST"),
+            Err(ApiError::InvalidWirePayload)
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliVerb::parse("create"),
+            Err(ApiError::InvalidWirePayload)
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliVerb::parse("status"),
+            Err(ApiError::InvalidWirePayload)
+        );
+    }
+
+    #[test]
+    fn from_args_refuses_empty_unknown_host_and_credential_flags() {
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(Vec::<String>::new(), "").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(["nope"], "").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(["list"], "").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(["list", "--host"], "").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(["list", "--host", "8.8.8.8:80"], "")
+                .unwrap_err(),
+            ApiError::AuthorizationDenied
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(["list", "--host", "not-a-socket"], "")
+                .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(
+                [
+                    "list",
+                    "--host",
+                    "127.0.0.1:18081",
+                    "--authorization",
+                    "secret"
+                ],
+                ""
+            )
+            .unwrap_err(),
+            ApiError::AuthorizationDenied
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(
+                ["list", "--host", "127.0.0.1:18081", "--pretty"],
+                ""
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(
+                ["list", "--host", "127.0.0.1:18081", "extra"],
+                ""
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(
+                [
+                    "list",
+                    "--host",
+                    "127.0.0.1:18081",
+                    "--run-id",
+                    "tepp-run-1"
+                ],
+                ""
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(
+                ["list", "--host", "127.0.0.1:18081"],
+                "{}"
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(
+                ["list", "--host", "127.0.0.1:18081", "--host", "127.0.0.1:9"],
+                ""
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(
+                ["list", "--host", "127.0.0.1:18081", "--consumer", "other"],
+                ""
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(
+                ["list", "--host", "127.0.0.1:18081", "--page-limit", "0"],
+                ""
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(
+                [
+                    "list",
+                    "--host",
+                    "127.0.0.1:18081",
+                    "--page-limit",
+                    &(ANALYSIS_RUN_COLLECTION_MAX_LIMIT + 1).to_string()
+                ],
+                ""
+            )
+            .unwrap_err(),
+            ApiError::LimitExceeded
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(
+                ["list", "--host", "127.0.0.1:18081", "--page-limit", "two"],
+                ""
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            AnalysisRunCollectionCliInvocation::from_args(
+                ["list", "--host", "127.0.0.1:18081", "--page-cursor", ""],
+                ""
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+    }
+
+    #[test]
+    fn list_assembles_default_consumer_and_optional_page_headers() {
+        let list = list_invocation();
+        assert_eq!(list.verb, AnalysisRunCollectionCliVerb::List);
+        assert_eq!(list.consumer, NARUON_CONSUMER_CODE);
+        assert!(list.page_cursor.is_none());
+        assert!(list.page_limit.is_none());
+        let http = compose_analysis_run_collection_cli_http(&list).expect("http");
+        assert!(http.starts_with("GET /v1/analysis-runs HTTP/1.1"));
+        assert!(http.contains("tepp-consumer: naruon"));
+        assert!(!http.contains("idempotency-key"));
+        assert!(!http.contains("tepp-page-cursor"));
+        assert!(!http.contains("tepp-page-limit"));
+
+        let paged = AnalysisRunCollectionCliInvocation::from_args(
+            [
+                "list",
+                "--host",
+                "127.0.0.1:18081",
+                "--consumer",
+                LINEAGEWEAVE_CONSUMER_CODE,
+                "--page-cursor",
+                "tepp-run-1",
+                "--page-limit",
+                "8",
+            ],
+            "",
+        )
+        .expect("paged");
+        assert_eq!(paged.consumer, LINEAGEWEAVE_CONSUMER_CODE);
+        let paged_http = compose_analysis_run_collection_cli_http(&paged).expect("paged http");
+        assert!(paged_http.contains("tepp-page-cursor: tepp-run-1"));
+        assert!(paged_http.contains("tepp-page-limit: 8"));
+        assert!(paged_http.contains("tepp-consumer: lineageweave"));
+    }
+
+    #[test]
+    fn dispatch_lists_created_and_cancelled_runs_without_scientific_acceptance() {
+        let mut service = AnalysisRunLiveService::new();
+        let first = request("cli-collection-idem-1");
+        let created = service.handle_http_request(&create_http(
+            &first,
+            NARUON_CONSUMER_CODE,
+            "127.0.0.1:18081",
+        ));
+        assert_eq!(created.status_code, 202);
+        let accepted = AnalysisRunAccepted::from_json(&created.body).expect("accepted");
+
+        let empty_before_second =
+            dispatch_analysis_run_collection_cli(&mut service, &list_invocation())
+                .expect("list one");
+        assert_eq!(empty_before_second.status_code, 200);
+        let listed =
+            render_analysis_run_collection_cli_stdout(&list_invocation(), &empty_before_second)
+                .expect("stdout");
+        assert!(!listed.contains(SCIENTIFIC_ACCEPTANCE_SCHEMA));
+        assert!(!listed.contains("rmse"));
+        assert!(!listed.contains("terminal_result"));
+        let page = AnalysisRunCollection::from_json(&listed).expect("page");
+        assert_eq!(page.runs.len(), 1);
+        assert_eq!(page.runs[0].run_id, accepted.run_id);
+        assert_eq!(
+            page.runs[0].run_state,
+            crate::AnalysisRunStatusState::Accepted
+        );
+
+        let second = request("cli-collection-idem-2");
+        let created_second = service.handle_http_request(&create_http(
+            &second,
+            NARUON_CONSUMER_CODE,
+            "127.0.0.1:18081",
+        ));
+        let accepted_second =
+            AnalysisRunAccepted::from_json(&created_second.body).expect("accepted second");
+        let cancel = format!(
+            "POST {NARUON_ANALYSIS_RUN_PATH}/{}/cancel HTTP/1.1\r\nHost: 127.0.0.1:18081\r\ncontent-type: application/json\r\ntepp-consumer: {NARUON_CONSUMER_CODE}\r\ntepp-contract-version: 1\r\nidempotency-key: {}\r\ncontent-length: 0\r\n\r\n",
+            accepted_second.run_id, second.idempotency_key
+        );
+        let cancelled = service.handle_http_request(&cancel);
+        assert_eq!(cancelled.status_code, 200);
+
+        let paged = AnalysisRunCollectionCliInvocation::from_args(
+            ["list", "--host", "127.0.0.1:18081", "--page-limit", "1"],
+            "",
+        )
+        .expect("limit 1");
+        let first_page =
+            dispatch_analysis_run_collection_cli(&mut service, &paged).expect("page 1");
+        let first_json =
+            render_analysis_run_collection_cli_stdout(&paged, &first_page).expect("page 1 stdout");
+        let first_collection = AnalysisRunCollection::from_json(&first_json).expect("first page");
+        assert_eq!(first_collection.runs.len(), 1);
+        let cursor = first_collection.next_cursor.expect("cursor");
+        let second_page_invocation = AnalysisRunCollectionCliInvocation::from_args(
+            [
+                "list",
+                "--host",
+                "127.0.0.1:18081",
+                "--page-cursor",
+                cursor.as_str(),
+                "--page-limit",
+                "1",
+            ],
+            "",
+        )
+        .expect("page 2");
+        let second_page =
+            dispatch_analysis_run_collection_cli(&mut service, &second_page_invocation)
+                .expect("page 2");
+        let second_json =
+            render_analysis_run_collection_cli_stdout(&second_page_invocation, &second_page)
+                .expect("page 2 stdout");
+        let second_collection =
+            AnalysisRunCollection::from_json(&second_json).expect("second page");
+        assert_eq!(second_collection.runs.len(), 1);
+        assert!(
+            second_collection
+                .runs
+                .iter()
+                .any(|row| row.run_state == crate::AnalysisRunStatusState::Cancelled)
+                || first_collection
+                    .runs
+                    .iter()
+                    .any(|row| row.run_state == crate::AnalysisRunStatusState::Cancelled)
+        );
+
+        let other = AnalysisRunCollectionCliInvocation::from_args(
+            [
+                "list",
+                "--host",
+                "127.0.0.1:18081",
+                "--consumer",
+                LINEAGEWEAVE_CONSUMER_CODE,
+            ],
+            "",
+        )
+        .expect("other consumer");
+        let isolated =
+            dispatch_analysis_run_collection_cli(&mut service, &other).expect("isolated");
+        let isolated_stdout =
+            render_analysis_run_collection_cli_stdout(&other, &isolated).expect("isolated stdout");
+        let isolated_page = AnalysisRunCollection::from_json(&isolated_stdout).expect("empty");
+        assert!(isolated_page.runs.is_empty());
+    }
+
+    #[test]
+    fn render_refuses_metrics_scientific_acceptance_and_empty_bodies() {
+        let list = list_invocation();
+        assert_eq!(
+            render_analysis_run_collection_cli_stdout(
+                &list,
+                &NaruonLiveResponse {
+                    status_code: 200,
+                    reason_phrase: "OK",
+                    body: String::new(),
+                }
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            render_analysis_run_collection_cli_stdout(
+                &list,
+                &NaruonLiveResponse {
+                    status_code: 200,
+                    reason_phrase: "OK",
+                    body: r#"{"contract_version":1,"runs":[],"rmse":1.0}"#.into(),
+                }
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            render_analysis_run_collection_cli_stdout(
+                &list,
+                &NaruonLiveResponse {
+                    status_code: 200,
+                    reason_phrase: "OK",
+                    body: format!(
+                        r#"{{"contract_version":1,"runs":[],"schema_version":"{SCIENTIFIC_ACCEPTANCE_SCHEMA}"}}"#
+                    ),
+                }
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        let error_stdout = render_analysis_run_collection_cli_stdout(
+            &list,
+            &NaruonLiveResponse {
+                status_code: 400,
+                reason_phrase: "Bad Request",
+                body: r#"{"error_code":"invalid_wire_payload"}"#.into(),
+            },
+        )
+        .expect("error");
+        assert!(error_stdout.contains("invalid_wire_payload"));
+        assert_eq!(
+            render_analysis_run_collection_cli_stdout(
+                &list,
+                &NaruonLiveResponse {
+                    status_code: 400,
+                    reason_phrase: "Bad Request",
+                    body: format!(r#"{{"schema_version":"{SCIENTIFIC_ACCEPTANCE_SCHEMA}"}}"#),
+                }
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            render_analysis_run_collection_cli_stdout(
+                &list,
+                &NaruonLiveResponse {
+                    status_code: 400,
+                    reason_phrase: "Bad Request",
+                    body: r#"{"rmse":0.1}"#.into(),
+                }
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        let empty_ok = render_analysis_run_collection_cli_stdout(
+            &list,
+            &NaruonLiveResponse {
+                status_code: 200,
+                reason_phrase: "OK",
+                body: r#"{"contract_version":1,"runs":[]}"#.into(),
+            },
+        )
+        .expect("empty");
+        assert!(empty_ok.contains("\"runs\":[]"));
+    }
+
+    #[test]
+    fn execute_over_tcp_and_parse_response_failures() {
+        let mut service = AnalysisRunLiveService::bind_loopback().expect("bind");
+        let addr = service.local_addr().expect("addr");
+        let handle = std::thread::spawn(move || {
+            drop(service.serve_one());
+        });
+        let mut invocation = list_invocation();
+        invocation.host = addr.to_string();
+        let response = execute_analysis_run_collection_cli(&invocation).expect("tcp");
+        assert_eq!(response.status_code, 200);
+        handle.join().expect("join");
+
+        invocation.host = "127.0.0.1:1".into();
+        assert_eq!(
+            execute_analysis_run_collection_cli(&invocation).unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+
+        let parsed =
+            parse_http_response(b"HTTP/1.1 200 OK\r\ncontent-length: 2\r\n\r\n{}").expect("parse");
+        assert_eq!(parsed.status_code, 200);
+        assert_eq!(
+            parse_http_response(b"not-http").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            parse_http_response(b"HTTP/1.0 200 OK\r\ncontent-length: 2\r\n\r\n{}").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            parse_http_response(b"HTTP/1.1 299 Mystery\r\ncontent-length: 2\r\n\r\n{}")
+                .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            parse_http_response(
+                b"HTTP/1.1 200 OK\r\ncontent-length: 2\r\ncontent-length: 2\r\n\r\n{}"
+            )
+            .unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            parse_http_response(b"HTTP/1.1 200 OK\r\ncontent-length: 9\r\n\r\n{}").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            parse_http_response(b"HTTP/1.1 200 OK\r\nbad-header\r\n\r\n{}").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            parse_http_response(&[0xff, 0xfe]).unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(static_reason(200).expect("200"), "OK");
+        assert_eq!(static_reason(202).expect("202"), "Accepted");
+        assert_eq!(static_reason(400).expect("400"), "Bad Request");
+        assert_eq!(static_reason(403).expect("403"), "Forbidden");
+        assert_eq!(static_reason(413).expect("413"), "Payload Too Large");
+        assert_eq!(static_reason(422).expect("422"), "Unprocessable Entity");
+        assert_eq!(
+            static_reason(500).unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            parse_http_response(b"HTTP/1.1\r\ncontent-length: 0\r\n\r\n").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            parse_http_response(b"HTTP/1.1 abc OK\r\ncontent-length: 0\r\n\r\n").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            parse_http_response(b"HTTP/1.1 200 OK\r\ncontent-length: x\r\n\r\n").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+        assert_eq!(
+            parse_http_response(b"HTTP/1.1 200 OK\r\nhost: 127.0.0.1\r\n\r\n{}").unwrap_err(),
+            ApiError::InvalidWirePayload
+        );
+    }
+
+    #[test]
+    fn stdin_reader_skips_terminal_and_reads_otherwise() {
+        let empty = read_analysis_run_collection_cli_stdin(true, std::io::empty()).expect("tty");
+        assert!(empty.is_empty());
+        let piped =
+            read_analysis_run_collection_cli_stdin(false, std::io::Cursor::new(b"leftover"))
+                .expect("piped");
+        assert_eq!(piped, "leftover");
+        let piped_empty = read_analysis_run_collection_cli_stdin(false, std::io::Cursor::new(b""))
+            .expect("empty");
+        assert!(piped_empty.is_empty());
+    }
+}
diff --git a/crates/tepp_api/src/bin/tepp_analysis_runs.rs b/crates/tepp_api/src/bin/tepp_analysis_runs.rs
new file mode 100644
index 000000000..87682715b
--- /dev/null
+++ b/crates/tepp_api/src/bin/tepp_analysis_runs.rs
@@ -0,0 +1,36 @@
+//! Operator CLI for loopback analysis-run collection GET.
+
+use std::io::{self, IsTerminal};
+use std::process::ExitCode;
+
+use tepp_api::{
+    AnalysisRunCollectionCliInvocation, AnalysisRunCollectionCliVerb, ApiError,
+    execute_analysis_run_collection_cli, read_analysis_run_collection_cli_stdin,
+    render_analysis_run_collection_cli_stdout,
+};
+
+fn main() -> ExitCode {
+    match run() {
+        Ok(()) => ExitCode::SUCCESS,
+        Err(_) => ExitCode::FAILURE,
+    }
+}
+
+fn run() -> Result<(), ApiError> {
+    let args: Vec<String> = std::env::args().skip(1).collect();
+    let verb =
+        AnalysisRunCollectionCliVerb::parse(args.first().ok_or(ApiError::InvalidWirePayload)?)?;
+    let body = read_analysis_run_collection_cli_stdin(io::stdin().is_terminal(), io::stdin())?;
+    let invocation = AnalysisRunCollectionCliInvocation::from_args(&args, body)?;
+    if invocation.verb != verb {
+        return Err(ApiError::InvalidWirePayload);
+    }
+    let response = execute_analysis_run_collection_cli(&invocation)?;
+    let stdout = render_analysis_run_collection_cli_stdout(&invocation, &response)?;
+    println!("{stdout}");
+    if (200..300).contains(&response.status_code) {
+        Ok(())
+    } else {
+        Err(ApiError::InvalidWirePayload)
+    }
+}
diff --git a/crates/tepp_api/src/lib.rs b/crates/tepp_api/src/lib.rs
index 2acf87810..ed5df35e7 100644
--- a/crates/tepp_api/src/lib.rs
+++ b/crates/tepp_api/src/lib.rs
@@ -14,6 +14,7 @@
 mod analysis_result;
 mod analysis_run;
 mod analysis_run_cancel_http;
+mod analysis_run_collection_cli;
 mod analysis_run_collection_http;
 mod analysis_run_live;
 mod analysis_run_status_http;
@@ -81,6 +82,20 @@ pub use analysis_run_cancel_http::AnalysisRunCancelRequest;
 pub use analysis_run_cancel_http::naruon_analysis_run_cancel_exchange;
 /// Refuse scientific-metric keys on a cancel payload.
 pub use analysis_run_cancel_http::refuse_metrics_on_cancel_payload;
+/// One validated collection CLI invocation.
+pub use analysis_run_collection_cli::AnalysisRunCollectionCliInvocation;
+/// Loopback collection CLI verb.
+pub use analysis_run_collection_cli::AnalysisRunCollectionCliVerb;
+/// Compose loopback collection GET bytes for a CLI invocation.
+pub use analysis_run_collection_cli::compose_analysis_run_collection_cli_http;
+/// Dispatch a collection CLI invocation against an in-process listener.
+pub use analysis_run_collection_cli::dispatch_analysis_run_collection_cli;
+/// Execute a collection CLI invocation over loopback TCP.
+pub use analysis_run_collection_cli::execute_analysis_run_collection_cli;
+/// Read leftover stdin for the collection CLI.
+pub use analysis_run_collection_cli::read_analysis_run_collection_cli_stdin;
+/// Render metric-free collection CLI stdout.
+pub use analysis_run_collection_cli::render_analysis_run_collection_cli_stdout;
 /// Analysis-run collection contract version constant.
 pub use analysis_run_collection_http::ANALYSIS_RUN_COLLECTION_CONTRACT_VERSION;
 /// Maximum exclusive collection cursor length.
diff --git a/crates/tepp_api/tests/analysis_run_collection_cli_contract.rs b/crates/tepp_api/tests/analysis_run_collection_cli_contract.rs
new file mode 100644
index 000000000..25da1c764
--- /dev/null
+++ b/crates/tepp_api/tests/analysis_run_collection_cli_contract.rs
@@ -0,0 +1,49 @@
+//! Contract tests for the analysis-run collection loopback CLI.
+
+use tepp_api::{
+    ANALYSIS_RUN_COLLECTION_CONTRACT_VERSION, AnalysisRunCollection,
+    AnalysisRunCollectionCliInvocation, AnalysisRunCollectionCliVerb, ApiError,
+    NARUON_CONSUMER_CODE, compose_analysis_run_collection_cli_http,
+};
+
+#[test]
+fn collection_cli_list_is_metric_free_get_without_credentials() {
+    assert_eq!(
+        AnalysisRunCollectionCliVerb::parse("list").expect("list"),
+        AnalysisRunCollectionCliVerb::List
+    );
+    let invocation =
+        AnalysisRunCollectionCliInvocation::from_args(["list", "--host", "127.0.0.1:18081"], "")
+            .expect("invocation");
+    assert_eq!(invocation.consumer, NARUON_CONSUMER_CODE);
+    let http = compose_analysis_run_collection_cli_http(&invocation).expect("http");
+    assert!(http.starts_with("GET /v1/analysis-runs HTTP/1.1"));
+    assert!(!http.contains("authorization"));
+    assert!(!http.contains("idempotency-key"));
+    assert!(!http.contains("copilot"));
+    assert_eq!(
+        ANALYSIS_RUN_COLLECTION_CONTRACT_VERSION,
+        AnalysisRunCollection::new(Vec::new(), None)
+            .expect("empty")
+            .contract_version
+    );
+}
+
+#[test]
+fn collection_cli_refuses_non_loopback_unknown_verbs_and_metric_bodies() {
+    assert_eq!(
+        AnalysisRunCollectionCliInvocation::from_args(["list", "--host", "8.8.8.8:80"], ""),
+        Err(ApiError::AuthorizationDenied)
+    );
+    assert_eq!(
+        AnalysisRunCollectionCliVerb::parse("create"),
+        Err(ApiError::InvalidWirePayload)
+    );
+    assert_eq!(
+        AnalysisRunCollectionCliInvocation::from_args(
+            ["list", "--host", "127.0.0.1:18081"],
+            r#"{"rmse":1.0}"#
+        ),
+        Err(ApiError::InvalidWirePayload)
+    );
+}
diff --git a/docs/API_CONTRACT.md b/docs/API_CONTRACT.md
index 88782097a..f9304407e 100644
--- a/docs/API_CONTRACT.md
+++ b/docs/API_CONTRACT.md
@@ -85,7 +85,9 @@ succeeded, failed, and unknown runs fail closed. `GET /v1/analysis-runs` on
 the loopback listener returns a metric-free collection of those states so
 operators do not guess run identities. Collection bodies never carry
 `tepp.scientific_acceptance.v1`. GET-by-id remains a later slice on this
-protected-main lineage.
+protected-main lineage. The loopback `tepp-analysis-runs list` CLI is the
+operator-visible client for that collection GET; it does not duplicate the
+scientific-acceptance `tepp-analysis-run` CLI.
 
 The stacked `analysis_engine` slice provides the first executable service-side
 path behind these DTOs. It consumes a bounded identity-free snapshot, excludes
diff --git a/docs/TRACEABILITY.md b/docs/TRACEABILITY.md
index b4509e6fd..5840d5fd6 100644
--- a/docs/TRACEABILITY.md
+++ b/docs/TRACEABILITY.md
@@ -55,6 +55,7 @@ The full APA 7th standards/literature register remains `docs/research/standards-
 | versioned service/API contracts and exports | PRD; API contract; ADR 0011/0013 | `tepp_api` analysis-run/export/JSON-LD/GraphML contracts on protected main (PR #21); request-bound terminal result active in PR #157; HTTP service remains accepted-target; the `orchestrator_live` loopback interpretation listener is on this PR | partial |
 | loopback analysis-run cancel HTTP | ADR 0029; API contract; RFC 9110 | `tepp_api` `POST /v1/analysis-runs/{run_id}/cancel` on `AnalysisRunLiveService`: metric-free cancelled status for accepted/running runs; succeeded/failed/unknown refuse; GET status remains a later slice | active-PR |
 | loopback analysis-run collection GET | ADR 0031; API contract; RFC 9110 | `tepp_api` `GET /v1/analysis-runs` on `AnalysisRunLiveService`: metric-free enumeration of accepted/running/cancelled/terminal runs; collection bodies refuse scientific-acceptance and RMSE keys; GET-by-id remains a later slice | active-PR |
+| loopback analysis-run collection CLI | ADR 0032; API contract; RFC 9110 | `tepp_api` `tepp-analysis-runs list` CLI: operator-visible metric-free collection client; `tepp.scientific_acceptance.v1` never prints; not GET-by-id and not the scientific-acceptance CLI | active-PR |
 | executable cutoff-safe analysis-run readiness | ADR 0021; temporal research; API terminal-result contract | stacked `analysis_engine` PR on #157: availability cutoff, snapshot binding, multiple-membership aggregation, digest-bound artifact, realistic end-to-end tests | active-PR |
 | delayed-reporting cutoff eligibility in truth corpora | ADR 0002; research | `tepp_simulation` eligible-at-cutoff filter on the active PR | active-PR |
 | versioned service/API contracts and exports | PRD; API contract; ADR 0011/0013 | `tepp_api` analysis-run/export/JSON-LD/GraphML contracts on protected main (PR #21); HTTP service remaining accepted-target | partial |
diff --git a/docs/adr/0032-analysis-run-collection-cli.md b/docs/adr/0032-analysis-run-collection-cli.md
new file mode 100644
index 000000000..642605492
--- /dev/null
+++ b/docs/adr/0032-analysis-run-collection-cli.md
@@ -0,0 +1,70 @@
+# ADR 0032 — Analysis-run collection loopback CLI
+
+**Decision status:** Accepted
+**Implementation maturity:** active-PR
+**Date:** 2026-08-31
+**Supersedes:** None; complements ADR 0031 for the operator-visible collection client. Does not supersede ADR 0014 claim-promotion authority. ADR 0026–0031 remain on live GAP-003A engine-library, terminal-wire DTO, GET-by-id, lifecycle-POST, cancel, scientific-acceptance CLI, and collection-GET slices.
+
+## Context
+
+ADR 0031 serves `GET /v1/analysis-runs` on the loopback listener, but operators still had to write raw HTTP/1.1 to enumerate accepted, running, cancelled, or terminal runs. Duplicating the collection GET listener, GET-by-id, lifecycle POST, cancel HTTP, or the scientific-acceptance CLI (`tepp-analysis-run` on live #362) would collide with live PRs.
+
+## Decision
+
+`tepp_api` publishes a loopback-only `tepp-analysis-runs` CLI:
+
+- `list` GETs `/v1/analysis-runs` with optional `tepp-page-cursor` / `tepp-page-limit`.
+- Stdout is the metric-free collection page: `run_id`, `run_state`, `idempotency_key`.
+- `tepp.scientific_acceptance.v1`, RMSE, bias, coverage, SE-gate, report, and `terminal_result` never appear.
+- Non-loopback hosts, unpublished consumers, credential-shaped flags, GET-by-id flags, nonempty stdin, unknown verbs, and hostile pagination fail closed.
+- Persistence, Compose recovery, and psychometric execution remain GAP-003B.
+
+## Alternatives considered
+
+1. **Keep raw HTTP as the only collection path** — rejected because operators still guess framing after ADR 0031.
+2. **Add `list` onto the live scientific-acceptance CLI (#362)** — rejected because that head already owns create/running/terminal/status and is stacked on GET-by-id, not collection GET.
+3. **Persist listed rows in PostgreSQL** — rejected as GAP-003B / live draft #287.
+4. **Loopback collection CLI with the same metric-free gates as ADR 0031** — accepted.
+
+## Consequences
+
+- Operators can enumerate runs on the same loopback listener that created them without writing HTTP.
+- Collection pages cannot be mistaken for a succeeded scientific-acceptance result.
+- CLI success is not release evidence.
+
+## Failure and recovery
+
+Non-loopback hosts return authorization denied. Unknown verbs, metric keys, nonempty bodies, unknown cursors, zero or non-integer limits, unpublished consumers, and credential flags fail closed. The in-memory registry is not durable.
+
+## Security, privacy, scientific-integrity, and governance impact
+
+- No credential headers cross the consumer boundary.
+- The CLI remains loopback-only and size-bounded.
+- Process exit 0 on a collection page is not measurement evidence and is not an ADR 0014 claim.
+
+## Compatibility and migration
+
+Collection GET, create POST, cancel POST, temporal-context, and project-history paths are unchanged. The scientific-acceptance CLI binary name `tepp-analysis-run` remains owned by ADR 0030 / #362. Production adapters may replace loopback while preserving metric-free collection rows.
+
+## Verification
+
+Falsifiable evidence:
+
+- CLI list JSON has no RMSE/bias/coverage/SE-gate/scientific-acceptance/`terminal_result` keys;
+- CLI list returns accepted and cancelled rows for one consumer and does not leak another consumer's runs;
+- non-loopback host, credential flags, GET-by-id flags, nonempty stdin, and unknown verbs fail closed;
+- Clippy `-D warnings`, `tepp_api` tests, rustdoc, and exact-head review remain required.
+
+## Rollback and supersession
+
+Rollback removes the `tepp-analysis-runs` binary and client module; collection GET remains valid. A superseding ADR is required to persist the registry, bind a public address, emit scientific-acceptance on the list, or treat CLI success as an ADR 0014 claim.
+
+## Related authority
+
+- ADR 0031 owns loopback collection GET.
+- ADR 0030 owns the scientific-acceptance loopback CLI (live #362).
+- ADR 0029 owns loopback cancel.
+- ADR 0018 owns consumer-scoped ingress and metric-free `202 Accepted`.
+- ADR 0014 owns scientific claim promotion.
+- ADR 0011 owns standalone/modular HTTP boundaries.
+- RFC 9110 owns GET semantics (Fielding, Nottingham, & Reschke, 2022). It does not authorize scientific claims.
diff --git a/docs/adr/README.md b/docs/adr/README.md
index b4111e69c..82318039c 100644
--- a/docs/adr/README.md
+++ b/docs/adr/README.md
@@ -32,6 +32,7 @@ Read [`ADR_POLICY.md`](ADR_POLICY.md) first. **Decision status and implementatio
 | [0025](0025-macos-native-rust-mlx-metal-boundary.md) | macOS-native Rust-owned MLX Metal execution | Accepted | accepted-target | Compose authenticates to a native host service; Linux never claims Metal, and actual backend/parity receipts fail closed. |
 | [0029](0029-analysis-run-cancel-http.md) | Loopback POST analysis-run cancel is metric-free cancelled status | Accepted | active-PR | Complements ADR 0018; does not supersede ADR 0014. ADR 0026–0028 live on other GAP-003A PRs. |
 | [0031](0031-analysis-run-collection-get.md) | Loopback GET analysis-run collection is metric-free enumeration | Accepted | active-PR | Complements ADR 0018/0029; does not supersede ADR 0014. ADR 0026–0030 live on other GAP-003A PRs. |
+| [0032](0032-analysis-run-collection-cli.md) | Loopback `tepp-analysis-runs list` is metric-free collection client | Accepted | active-PR | Complements ADR 0031; does not supersede ADR 0014. ADR 0026–0031 live on other GAP-003A PRs. |
 | [0023](0023-lineage-criterion-anchor-contract.md) | TEPP-owned Event Lineage criterion anchor | Accepted | active-PR | PR #237 publishes the strict accepted/rejected artifact and identities; estimator execution remains fail-closed future work. |
 | [0024](0024-independent-topic-importance-anchor.md) | Posterior topic-context producer contract | Accepted | contract-only active-PR | Strict DTO/schema only; the current estimator does not emit it. fast-mlsirm owns case-deletion influence. |
 | [0001](0001-rust-first-modular-msa.md) | Rust-first numerical core and CPU `f64` reference | Accepted | partial | ADR 0011 owns cross-service/MSA authority; 0001 retains numerical/backend authority. |
@@ -144,6 +145,7 @@ Use the narrowest owning ADR when decisions overlap:
 - **macOS-native Rust-owned MLX Metal execution:** ADR 0024.
 - **analysis-run cancel HTTP:** ADR 0029.
 - **analysis-run collection GET:** ADR 0031.
+- **analysis-run collection CLI:** ADR 0032.
 
 ## Change and supersession rule
 
diff --git a/docs/connectors/naruon-artifact-consumer.md b/docs/connectors/naruon-artifact-consumer.md
index 5bdc328a1..0bdd58d9d 100644
--- a/docs/connectors/naruon-artifact-consumer.md
+++ b/docs/connectors/naruon-artifact-consumer.md
@@ -27,6 +27,7 @@ TEPP remains the scientific authority for estimation, recovery metrics, temporal
 | purpose-bound export auth | `tepp_api` `authorize_export` with `ModularServiceConsumer` | TEPP gate |
 | HTTP analysis-run create | `tepp_api` `naruon_analysis_run_exchange` → `POST /v1/analysis-runs` | naruon → TEPP |
 | HTTP analysis-run collection | `tepp_api` `naruon_analysis_run_collection_exchange` → `GET /v1/analysis-runs` | naruon → TEPP |
+| CLI analysis-run collection | `tepp_api` `tepp-analysis-runs list` → loopback `GET /v1/analysis-runs` | naruon → TEPP |
 | HTTP analysis-run cancel | `tepp_api` `naruon_analysis_run_cancel_exchange` → `POST /v1/analysis-runs/{run_id}/cancel` | naruon → TEPP |
 | HTTP export authorize | `tepp_api` `naruon_export_exchange` → `POST /v1/exports` | naruon → TEPP |
 | Live loopback POST | `tepp_api` `NaruonLiveService` → `POST /v1/analysis-runs` and `/v1/exports` | naruon → TEPP |
diff --git a/docs/research/analysis-run-collection-cli.md b/docs/research/analysis-run-collection-cli.md
new file mode 100644
index 000000000..30b281dc5
--- /dev/null
+++ b/docs/research/analysis-run-collection-cli.md
@@ -0,0 +1,56 @@
+# Analysis-run collection CLI (doctoring)
+
+## Scope
+
+`tepp-analysis-runs list` is the operator-visible client of loopback
+`GET /v1/analysis-runs`. HTTP method, path, and header semantics follow current
+HTTP semantics (Fielding, Nottingham, & Reschke, 2022). Fail-closed refusal of
+non-loopback hosts, unpublished consumers, review/Copilot/GitHub credential
+flags, and scientific-authority promotion is repository contract authority
+(ADR 0032; ADR 0031; ADR 0018; ADR 0011), not an RFC inference rule.
+
+CLI stdout is metric-free `AnalysisRunCollection` JSON. Each row carries
+`run_id`, `run_state`, and `idempotency_key` only. Process exit 0 is not a
+completed temporal model, calibrated score, theta estimate, uncertainty
+statement, or scientific claim. `tepp.scientific_acceptance.v1` never appears.
+
+## Authority
+
+### External standards (HTTP only)
+
+Fielding, R., Nottingham, M., & Reschke, J. (Eds.). (2022). *HTTP semantics*
+(RFC 9110). IETF. https://doi.org/10.17487/RFC9110
+
+RFC 9110 §9.3.1 describes GET as a method for retrieving the target resource's
+current state. TEPP maps that retrieval onto a bounded, consumer-scoped
+collection of metric-free run rows. The RFC does not define psychometric
+acceptance, RMSE, or claim promotion.
+
+### Internal contract evidence
+
+- `docs/adr/0032-analysis-run-collection-cli.md` — this client
+- `docs/adr/0031-analysis-run-collection-get.md` — collection GET listener
+- `docs/adr/0030-scientific-acceptance-loopback-cli.md` — distinct
+  `tepp-analysis-run` scientific-acceptance CLI on live #362
+- `docs/adr/0018-consumer-scoped-analysis-run-ingress.md` — closed consumer
+  registry and metric-free `202 Accepted`
+- `docs/adr/0014-scientific-claim-promotion-and-release-evidence.md` — CLI
+  success is not a scientific claim
+- `docs/API_CONTRACT.md` — documented collection resource
+- `crates/tepp_api/tests/analysis_run_collection_cli_contract.rs` —
+  fail-closed collection CLI proofs
+
+## Verification
+
+- `tepp-analysis-runs list` of accepted and cancelled runs returns metric-free
+  rows without RMSE/bias/coverage/SE-gate keys or `tepp.scientific_acceptance.v1`;
+- another consumer cannot read the first consumer's rows;
+- non-loopback hosts, credential flags, GET-by-id flags, nonempty stdin, and
+  unknown verbs fail closed;
+- review, Copilot, GitHub, and bearer flags remain `AuthorizationDenied`.
+
+## Non-claims
+
+This slice does not implement GET-by-id, running/terminal POST, cancel CLI,
+scientific-acceptance CLI verbs, persistence, production TLS, Leiden consensus,
+or an ADR 0014 scientific claim-promotion package.