test(reborn): annotate the §4.3 store ratchet with per-entry achievable-floor status

[ilblackdragon]

What

The mechanical §4.3 store consolidations are complete (A1–A8: approvals, authorization, processes, run-state, budget-gate, and the whole outbound family). Every entry still in FROZEN_INMEMORY_STORES is blocked on non-mechanical work OR is a justified keep — so the §10 "shrink to empty" goal is not reachable by a swap.

This annotates each remaining allowlist entry with its verified status so the next contributor picks up a scoped task instead of re-deriving the blocker. Comments are stripped by the scanner — documentation only, the enforced string set is unchanged (entries reordered to group the justified keeps).

Per-entry status

• turns cluster — DEFERRED (production inmemory-turn-state authority; needs a no-livelock concurrency proof + a built filesystem variant, some cross-crate).
• InMemoryBoundedSubagentGoalStore, InMemoryOpenAiCompatRefStore — JUSTIFIED bounded caches (capacity-bounded evict-oldest / bounded-LRU + filesystem-free contract boundary), NOT persistence debt; a durable variant would be semantically wrong.
• InMemoryExtensionInstallationStore — BLOCKED cross-crate (the Filesystem variant in composition depends on a composition-internal contract registry; can't move down to ironclaw_extensions).
• InMemorySecretStore — security-sensitive.
• InMemorySessionStore — BUILD-FIRST (no filesystem variant; auth-adjacent).

Also reconciles the module-doc "definition of done" to note the two justified caches (a future PR may formally split them into a justified-keep list).

Verification

No production code changes. Ratchet self-tests + the frozen-set contract still pass (4 tests); fmt + pre-commit clean.

Stack

Stacked on #6214. Closes out the §4.3 store-consolidation axis with documented evidence of the achievable floor.

🤖 Generated with Claude Code

[ironloopai]

🔎 IronLoop Review Status

Head: 5985c26f34925ec66185ad43d631395d015bc716
Result: One or more review results were superseded by a newer PR head.
Next: Run @ironloopai review on the latest PR head.
Updated: 2026-07-18T03:30:14.160Z

Current reviewers:

Reviewer	State	Verdict	Findings	Last update
ironloop/common-reviewer (reviewer)	Superseded	N/A	N/A	2026-07-18T01:07:23.253Z

Reviewer summaries

Reviewer	Detail
ironloop/common-reviewer (reviewer)	Superseded by a newer PR head. New head: 843cfba. Previous verdict: Approved.

Recent activity

Time	Reviewer	State	Detail
2026-07-18T00:31:07.676Z	ironloop/common-reviewer (reviewer)	Queued	Accepted review request for head cf8a845.
2026-07-18T00:31:07.676Z	ironloop/common-reviewer (reviewer)	Queued	Waiting for this reviewer lane to become available.
2026-07-18T00:31:08.589Z	ironloop/common-reviewer (reviewer)	Started	Reviewer worker started.
2026-07-18T00:31:11.013Z	ironloop/common-reviewer (reviewer)	Workspace ready	Prepared isolated checkout (merge_ref) at 16e5812.
2026-07-18T00:35:23.864Z	ironloop/common-reviewer (reviewer)	Result captured	Approved; 0 blocking findings.
2026-07-18T00:35:23.864Z	ironloop/common-reviewer (reviewer)	Completed	Review completed and terminal status was persisted.
2026-07-18T01:07:23.253Z	ironloop/common-reviewer (reviewer)	Superseded	A newer PR head replaced this review (843cfba).

Available commands

• @ironloopai help
• @ironloopai agents
• @ironloopai review
• @ironloopai review --agent <agent>

Run metadata

Admission: webhook accepted the request and IronLoop persisted reviewer state before this projection.

[coderabbitai]

📝 Walkthrough

Summary by CodeRabbit

• Documentation

  • Clarified the completion criteria for in-memory store tracking.
  • Added status details and rationale for listed bounded caches.
  • Updated the inventory and organization of the documented allowlist.

• Tests

  • No test behavior or validation logic was changed.

Walkthrough

The ratchet test’s contract comments now describe completion as empty debt, document bounded-cache exceptions and inventory status, and reorder entries in the frozen in-memory store list. Test logic and public declarations are unchanged.

Changes

In-memory store ratchet

Layer / File(s)	Summary
Ratchet contract and inventory documentation crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs	The definition-of-done wording and FROZEN_INMEMORY_STORES documentation now describe debt status, bounded-cache exceptions, triage notes, and deferred entries; the allowlist ordering was adjusted without changing its set of entries.

Estimated code review effort: 1 (Trivial) | ~3 minutes

Possibly related PRs

• nearai/ironclaw#6204: Modifies the same ratchet test and frozen in-memory store contract.
• nearai/ironclaw#6213: Updates the same allowlist and surrounding ratchet documentation.

🚥 Pre-merge checks | ✅ 4

✅ Passed checks (4 passed)

Check name	Status	Explanation
Title check	✅ Passed	Conventional Commits style is used and the title accurately matches the store-ratchet documentation change.
Description check	✅ Passed	The description is detailed and covers summary plus verification, but it omits several template sections like issue, security, and rollback.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[gemini-code-assist]

Code Review

This pull request updates the documentation and comments in reborn_inmemory_store_ratchet.rs to reflect the current status of the remaining in-memory stores after the mechanical consolidations (A1–A8) have been completed. It categorizes the remaining stores into deferred, justified keeps (bounded caches), blocked, security-sensitive, and build-first categories, providing clear context for future refactoring efforts. There are no review comments, so no feedback is provided.

Important

The consumer version of Gemini Code Assist on GitHub is being sunset. Starting June 18, 2026, new organization installations will be blocked, and all code review activity will officially cease on July 17, 2026.
For more details on the timeline and next steps, please review the Help Documentation.

[ironloopai]

✅ IronLoop Review: reviewer

Review at a glance

Verdict	Blocking	Notes	Inline	Head
✅ Approved	0	5	5	cf8a84512d6b

Head: cf8a84512d6b9cc7ae7409e119073e34047276ce
Next: No reviewer action needed.

Run details

Status: Current
Needs human: no
Needs validation: no

Summary

This is a bounded, reviewable stack layer changing one architecture-test documentation file (52 additions, 15 deletions). The enforced allowlist set is unchanged and no runtime behavior changes. Static checks passed, with five non-blocking accuracy/completeness issues in the new annotations.

Findings

Blocking: 0 / Notes: 5

Non-blocking notes (5)

1. 💬 [LOW] Referenced store-triage worklog is absent

Location: crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs:64
The target tree contains no .worklog/reborn-refactor.md, so future contributors cannot access the cited “full analysis.” Commit/link an available source of truth or remove this reference.

2. 💬 [LOW] CheckpointState filesystem variant already exists

Location: crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs:70-71
FilesystemCheckpointStateStore already exists in ironclaw_loop_host, has contract tests, and is wired in composition. Separate InMemoryCheckpointStateStore from the stores that genuinely still need a filesystem implementation so the next task is scoped correctly.

3. 💬 [LOW] Justified-cache rationale contradicts production filesystem stores

Location: crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs:76-79
The blanket claim that durable variants would be semantically wrong conflicts with the code: FilesystemSubagentGoalStore is explicitly the production store for libSQL/Postgres, and serving wires FilesystemOpenAiCompatRefStore. The in-memory implementations may remain justified test/fallback implementations, but the annotation should state that rationale instead.

4. 💬 [LOW] OpenAI compatibility store is not LRU

Location: crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs:83-85
The implementation evicts the mapping with the minimum created_at; lookups and mutations do not refresh recency. Describe it as bounded oldest-created eviction rather than LRU.

5. 💬 [LOW] Three remaining entries are explicitly untriaged

Location: crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs:102-107
This block says the three crate-private stores are “not yet individually triaged,” contradicting the preceding claim that every entry has a verified per-entry status and a scoped next task. Add a concrete status for each entry or soften the overall completeness claim.

Developer follow-up

After fixing this feedback:

1. Push the fix to this PR branch.
2. Re-run this reviewer with @ironloopai review --agent reviewer if you only changed this reviewer's findings.
3. Re-run all reviewers with @ironloopai review when the fix may affect multiple areas.

[ironloopai]

This referenced worklog is not present anywhere in the target tree, so future contributors cannot consult the claimed full analysis. Please link an available source of truth or remove the reference.

[ilblackdragon]

Fixed on this branch: the worklog reference is removed — the annotations themselves are now the source of truth, and the completeness claim was softened to match.

[ironloopai]

FilesystemCheckpointStateStore is already implemented in ironclaw_loop_host, contract-tested, and wired by composition. Please distinguish InMemoryCheckpointStateStore from the stores that still need a filesystem variant.

[ilblackdragon]

Fixed: the turns-cluster note now distinguishes InMemoryCheckpointStateStore (its FilesystemCheckpointStateStore already exists in ironclaw_loop_host, contract-tested and composition-wired — only the test-seam swap + allowlist trim remain) from LoopCheckpoint/InstructionMaterialization, which still need filesystem variants built.

[ironloopai]

This rationale conflicts with existing production adapters: libSQL/Postgres use FilesystemSubagentGoalStore, and OpenAI-compatible serving uses FilesystemOpenAiCompatRefStore. The in-memory types may be justified test/fallback implementations, but durable storage is not semantically wrong for these contracts.

[ilblackdragon]

Fixed: the justified-keeps section now states that durable production variants already exist and are wired (FilesystemSubagentGoalStore in the libSQL/Postgres runner adapters, FilesystemOpenAiCompatRefStore in OpenAI-compatible serving), and frames the in-memory types as the bounded volatile role beside those stores rather than missing consolidations.

[ironloopai]

This is not LRU: eviction selects the minimum created_at, and accesses do not update recency. Please describe it as oldest-created eviction.

[ilblackdragon]

Fixed: described as capacity-bounded with oldest-created eviction (evicts the minimum created_at; reads do not refresh recency — explicitly noted as NOT an LRU), matching refs.rs's min_by_key(created_at).

[ironloopai]

These entries are explicitly untriaged, which conflicts with the preceding claim that every remaining entry has a verified status and scoped task. Please annotate each one or soften the completeness claim.

[ilblackdragon]

Fixed: the header claim is softened — triaged entries carry their scoped task, and the pub(crate) trio is explicitly labeled untriaged instead of being covered by a blanket verified-status claim.

[railway-app]

🚅 Deployed to the ironclaw-pr-6216 environment in ironclaw-ci-preview

Service	Status	Web	Updated (UTC)
ironclaw	✅ Success (View Logs)	Web	Jul 18, 2026 at 3:30 am

[ilblackdragon]

✅ Ready for merge

CI green (18 pass / 0 fail) on the restacked head.

All five IronLoop factual findings fixed with replies: the missing-worklog reference removed; InMemoryCheckpointStateStore distinguished from stores needing variants built (its FilesystemCheckpointStateStore already exists, contract-tested and wired); the justified-keeps section acknowledges the production FilesystemSubagentGoalStore/FilesystemOpenAiCompatRefStore variants; eviction described as oldest-created (verified against refs.rs's min_by_key(created_at)), not LRU; the completeness claim softened to label the pub(crate) trio explicitly untriaged.

🤖 Generated with Claude Code

[ilblackdragon]

Post-restack confirmation: rebased onto the updated chain after the #6212/#6214/#6221 merges. CI fully green — 18 pass / 0 fail. Still ready for merge.

[coderabbitai]

Actionable comments posted: 2

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs`:
- Around line 26-35: The ratchet in the test around FROZEN_INMEMORY_STORES does
not detect newly added frozen stores. Add a checked-in baseline and compare the
current frozen set against it, failing when entries are added while allowing
removals; alternatively remove the “only shrinks” guarantee from the surrounding
comment and failure text.
- Around line 57-111: Add inline source anchors to the durable-variant and
wiring claims in FROZEN_INMEMORY_STORES, citing the defining and integration
symbols for FilesystemCheckpointStateStore, FilesystemSubagentGoalStore,
FilesystemOpenAiCompatRefStore, and FilesystemExtensionInstallationStore. Keep
FilesystemSessionStore described as an unimplemented open gap because no crates/
source exists, and trim any unsupported claims rather than leaving them uncited.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: ASSERTIVE

Plan: Pro Plus

Run ID: cf4571cc-557f-43ae-b73d-7747c87b4b45

📥 Commits

Reviewing files that changed from the base of the PR and between a900da5 and 5985c26.

📒 Files selected for processing (1)

• crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs

[coderabbitai]

🎯 Functional Correctness | 🟠 Major | 🏗️ Heavy lift

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
set -euo pipefail

git ls-files | rg '(^|/)reborn_inmemory_store_ratchet\.rs$|(^|/)CLAUDE\.md$|(^|/)AGENTS\.md$|(^|/)\.claude/rules'
printf '\n--- file outline ---\n'
ast-grep outline crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs --view expanded
printf '\n--- relevant file slice ---\n'
sed -n '1,240p' crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs

Repository: nearai/ironclaw

Length of output: 17702

---

Enforce the “only shrinks” ratchet.
added.is_empty() only rejects names outside the current allowlist. If a new InMemory*Store is added to both the code and FROZEN_INMEMORY_STORES, this test still passes, so the “never add” / “only shrinks” contract is not enforced. Add a checked-in baseline/diff check, or drop that guarantee from the comment and failure text.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs` around
lines 26 - 35, The ratchet in the test around FROZEN_INMEMORY_STORES does not
detect newly added frozen stores. Add a checked-in baseline and compare the
current frozen set against it, failing when entries are added while allowing
removals; alternatively remove the “only shrinks” guarantee from the surrounding
comment and failure text.

[coderabbitai]

📐 Maintainability & Code Quality | 🟡 Minor | ⚡ Quick win

🧩 Analysis chain

🏁 Script executed:

set -euo pipefail

printf '\n== File outline ==\n'
wc -l crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs
sed -n '1,220p' crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs

printf '\n== Search referenced symbols ==\n'
rg -n "FilesystemCheckpointStateStore|FilesystemSubagentGoalStore|FilesystemOpenAiCompatRefStore|FilesystemExtensionInstallationStore|InMemoryTurnStateStore|InMemoryCheckpointStateStore|InMemoryLoopCheckpointStore|InMemoryInstructionMaterializationStore|InMemoryBoundedSubagentGoalStore|InMemoryOpenAiCompatRefStore|InMemoryExtensionInstallationStore|InMemorySecretStore|InMemorySessionStore" crates -g '!target'

printf '\n== CLAUDE/AGENTS invariant mentions ==\n'
rg -n "load-bearing claims|Everything Goes Through Tools|LLM data is never deleted|Trusted-ingress seal|Test through the caller|filesystem variant|bounded in-memory caches|composition-internal contract registry|no-CAS-livelock|oldest-created eviction|FilesystemSessionStore|FilesystemExtensionInstallationStore" CLAUDE.md AGENTS.md .claude -g '!target'

Repository: nearai/ironclaw

Length of output: 50372

---

🏁 Script executed:

set -euo pipefail
sed -n '1,220p' crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs

Repository: nearai/ironclaw

Length of output: 11287

---

🏁 Script executed:

set -euo pipefail

printf '\n== FilesystemCheckpointStateStore definition/re-export ==\n'
rg -n "pub struct FilesystemCheckpointStateStore|pub use filesystem_checkpoint_state::FilesystemCheckpointStateStore" \
  crates/ironclaw_loop_host/src crates/ironclaw_loop_host/tests

printf '\n== FilesystemSubagentGoalStore definition/wiring ==\n'
rg -n "pub struct FilesystemSubagentGoalStore|FilesystemSubagentGoalStore::new|FilesystemSubagentGoalStore" \
  crates/ironclaw_runner crates/ironclaw_reborn_composition/src/runtime.rs crates/ironclaw_reborn_composition/tests

printf '\n== FilesystemOpenAiCompatRefStore definition/wiring ==\n'
rg -n "pub struct FilesystemOpenAiCompatRefStore|FilesystemOpenAiCompatRefStore::with_root|FilesystemOpenAiCompatRefStore" \
  crates/ironclaw_reborn_composition/src crates/ironclaw_reborn_composition/tests crates/ironclaw_reborn_openai_compat crates/ironclaw_reborn_composition/src/llm_admin

printf '\n== FilesystemExtensionInstallationStore definition/wiring ==\n'
rg -n "pub(crate) struct FilesystemExtensionInstallationStore|FilesystemExtensionInstallationStore::load_at|FilesystemExtensionInstallationStore::default_state_path|FilesystemExtensionInstallationStore" \
  crates/ironclaw_reborn_composition/src crates/ironclaw_reborn_composition/tests crates/ironclaw_extensions crates/ironclaw_reborn_composition/src/extension_host

printf '\n== Bounded-cache semantics ==\n'
rg -n "capacity-bounded|evict-oldest|oldest-created|LRU|reads do not refresh recency|minimum `created_at`|VecDeque insertion order" \
  crates/ironclaw_runner/src crates/ironclaw_reborn_openai_compat crates/ironclaw_reborn_composition/src crates/ironclaw_reborn_composition/tests

printf '\n== FilesystemSessionStore search ==\n'
rg -n "FilesystemSessionStore" crates

Repository: nearai/ironclaw

Length of output: 12381

---

Add source anchors to the ratchet notes The status block makes load-bearing claims about durable variants and wiring; cite the defining/writing sites inline, or trim the prose. For example: FilesystemCheckpointStateStore (crates/ironclaw_loop_host/src/filesystem_checkpoint_state.rs:44, re-exported in crates/ironclaw_loop_host/src/lib.rs:79), FilesystemSubagentGoalStore (crates/ironclaw_runner/src/subagent/goal_store.rs:79, wired in crates/ironclaw_reborn_composition/src/runtime.rs:305,368), FilesystemOpenAiCompatRefStore (crates/ironclaw_reborn_openai_compat/src/refs_storage.rs:37, wired in crates/ironclaw_reborn_composition/src/llm_admin/openai_compat_serve.rs:176), and FilesystemExtensionInstallationStore (crates/ironclaw_reborn_composition/src/extension_host/extension_installation_store.rs:19, loaded in crates/ironclaw_reborn_composition/src/factory.rs:1925,3326,3367). FilesystemSessionStore still has no crates/ hit, so keep that one framed as an open gap.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@crates/ironclaw_architecture/tests/reborn_inmemory_store_ratchet.rs` around
lines 57 - 111, Add inline source anchors to the durable-variant and wiring
claims in FROZEN_INMEMORY_STORES, citing the defining and integration symbols
for FilesystemCheckpointStateStore, FilesystemSubagentGoalStore,
FilesystemOpenAiCompatRefStore, and FilesystemExtensionInstallationStore. Keep
FilesystemSessionStore described as an unimplemented open gap because no crates/
source exists, and trim any unsupported claims rather than leaving them uncited.

Source: Coding guidelines

[github-actions]

Coverage ratchet

Ratchet mode: ENFORCING

RATCHET PASS: global
  observed: 85.59% (306535 / 358149 lines)
  floor:    85.3% (tolerance 0.5pp -> effective floor 84.8%)
  denominator: 358149 lines now vs 320188 at floor capture (+37961 lines, +11.86%) — material change (>5%)

⚠️ 2 Reborn crate(s) have 0 int-tier coverage (target: 0) — ironclaw_prompt_envelope, ironclaw_scripts

Reborn integration-tier coverage

Line coverage (Reborn crates): 85.59% — 306535 / 358149 lines

Per-crate breakdown (65 crates, lowest-covered first)

Crate	Line %	Covered / Total
ironclaw_prompt_envelope	0%	0 / 88
ironclaw_scripts	0%	0 / 345
ironclaw_runtime_policy	31.75%	80 / 252
ironclaw_event_projections	43.31%	673 / 1554
ironclaw_observability	61.54%	16 / 26
ironclaw_authorization	62.46%	604 / 967
ironclaw_mcp	64.89%	595 / 917
ironclaw_triggers	65.44%	2142 / 3273
ironclaw_dispatcher	67.15%	92 / 137
ironclaw_filesystem	67.69%	3932 / 5809
ironclaw_channel_host	68.65%	219 / 319
ironclaw_memory	69.2%	773 / 1117
ironclaw_reborn_migration	71.64%	1551 / 2165
ironclaw_trust	72.88%	661 / 907
ironclaw_reborn_cli	73.98%	7095 / 9591
ironclaw_capabilities	74.36%	1685 / 2266
ironclaw_wasm_limiter	74.6%	47 / 63
ironclaw_reborn_event_store	74.67%	958 / 1283
ironclaw_extractors	74.72%	538 / 720
ironclaw_projects	76.48%	400 / 523
ironclaw_llm	78.27%	20216 / 25827
ironclaw_product_context	78.57%	11 / 14
ironclaw_telegram_extension	80.18%	4842 / 6039
ironclaw_wasm_product_adapters	80.36%	1448 / 1802
ironclaw_process_sandbox	80.65%	671 / 832
ironclaw_first_party_extensions	81.06%	5965 / 7359
ironclaw_memory_native	81.22%	3205 / 3946
ironclaw_events	81.43%	1539 / 1890
ironclaw_secrets	82.79%	2794 / 3375
ironclaw_network	82.98%	673 / 811
ironclaw_reborn_identity	83.59%	433 / 518
ironclaw_processes	83.76%	939 / 1121
ironclaw_run_state	83.96%	424 / 505
ironclaw_reborn_config	84.02%	1830 / 2178
ironclaw_wasm	84.44%	1069 / 1266
ironclaw_auth	84.81%	3233 / 3812
ironclaw_product_workflow	84.91%	11031 / 12992
ironclaw_turns	85.13%	13737 / 16136
ironclaw_host_api	85.37%	2701 / 3164
ironclaw_channel_delivery	85.79%	1383 / 1612
ironclaw_common	86.13%	1714 / 1990
ironclaw_threads	86.93%	4708 / 5416
ironclaw_slack_v2_adapter	87.3%	1491 / 1708
ironclaw_skills	87.58%	4470 / 5104
ironclaw_hooks	87.78%	9921 / 11302
ironclaw_product_adapter_registry	88.06%	531 / 603
ironclaw_product_adapters	88.1%	3384 / 3841
ironclaw_reborn_traces	88.2%	11946 / 13544
ironclaw_reborn_composition	88.38%	70500 / 79769
ironclaw_webui	88.42%	7333 / 8293
ironclaw_host_runtime	88.76%	18005 / 20284
ironclaw_extensions	89.38%	2971 / 3324
ironclaw_runner	89.5%	16990 / 18983
ironclaw_reborn_openai_compat	89.5%	3778 / 4221
ironclaw_telegram_v2_adapter	89.7%	2717 / 3029
ironclaw_approvals	90.18%	1598 / 1772
ironclaw_conversations	90.39%	3123 / 3455
ironclaw_event_streams	90.82%	1009 / 1111
ironclaw_resources	91.65%	4476 / 4884
ironclaw_loop_host	92.25%	15051 / 16316
ironclaw_attachments	93.06%	630 / 677
ironclaw_agent_loop	94.88%	9184 / 9680
ironclaw_safety	95.04%	3677 / 3869
ironclaw_outbound	95.52%	3451 / 3613
ironclaw_first_party_extension_ports	95.62%	3672 / 3840

This table itself is informational and never gates the PR on its own — not the percentage, not the per-crate holes, not the 0-coverage callout. A separate coverage ratchet (dry-run until enforce=true; see tests/integration/coverage-floor.toml) can fail the build on specific configured floors.

Exemptions (3 entry/entries excluded from the accounting above)

Module / Crate	Reason	Issue
crate: ironclaw_embeddings	v1-only: consumed only by root ironclaw (src/app.rs, src/tools/builtin/memory.rs, src/workspace/mod.rs, src/config/{mod,embeddings}.rs); no crates/* dependents. Covered by "Tests (Legacy)".	#5657
crate: ironclaw_gateway	v1-only: consumed only by root ironclaw (src/channels/web/platform/static_files.rs, src/channels/web/handlers/frontend.rs); no crates/* dependents. Covered by "Tests (Legacy)".	#5657
crate: ironclaw_tui	v1-only: consumed only by root ironclaw (src/main.rs, src/channels/tui.rs); no crates/* dependents. Crate's own doc comment confirms it bridges INTO v1, not Reborn. Covered by "Tests (Legacy)".	#5657

[ilblackdragon]

Post-restack confirmation (after #6213's merge): rebased onto main, CI fully green on the current head. Still ready for merge.