fix(hook): stdin timeout + stderr logging + bare /agentflare report

[getappz]

Changes

• Add
  ead_stdin_timeout(ms)\ helper using thread+channel+recv_timeout (stdlib only)
• Replace blocking \stdin().read_to_string()\ in \pre_tool_use\ and \prompt_submit\
• 1s safety timeout, fail-open: timeout → log to stderr, skip gracefully
• Bare /agentflare\ or /agentflare status\ reports active state (on/off)

Tickets

• Closes ponytail PR audit [ponytail#214] Windows stdin timeout — hooks must never block on Windows #71 (Windows stdin timeout)
• Closes ponytail PR audit [ponytail#303] Log errors to stderr, never swallow silently #70 (stderr logging)
• Closes ponytail PR audit [ponytail#99] Bare /ponytail reports active mode, doesn't reset #75 (bare /agentflare report)

References

• ponytail#214 (upstream Windows stdin hang fix)
• ponytail#303 (upstream stderr logging)
• ponytail#99 (upstream bare /ponytail report)

Test

• 186/186 tests pass

Summary by CodeRabbit

• New Features

  • Added a status response for /agentflare and /agentflare status to report whether the hook is currently ACTIVE or off.

• Bug Fixes

  • Prevented indefinite waiting by adding stdin read timeouts for hook handlers.
  • Improved behavior when input is missing or empty by skipping processing and showing clearer feedback.

• Tests

  • Added a unit test to ensure the timeout-based stdin reader does not panic.

[coderabbitai]

📝 Walkthrough

Walkthrough

Adds a read_stdin_timeout helper in src/hook.rs that reads stdin with a 1-second timeout via a spawned thread, applies it in pre_tool_use and prompt_submit with early returns on timeout/empty input, adds /agentflare status command handling, and includes a unit test.

Changes

Hook stdin timeout and status command

Layer / File(s)	Summary
Stdin timeout helper src/hook.rs	Introduces read_stdin_timeout(ms) that spawns a thread to read stdin and returns None on timeout or empty input via recv_timeout.
Apply timeout reads in hook entry points src/hook.rs	Replaces blocking read_to_string calls in pre_tool_use and prompt_submit with read_stdin_timeout(1000), adding early returns with eprintln! messages on timeout/empty input.
/agentflare status command and test src/hook.rs	Adds /agentflare and /agentflare status handling that returns a JSON hookSpecificOutput reporting ACTIVE/off state, and adds a unit test verifying read_stdin_timeout does not panic.

Estimated code review effort: 2 (Simple) | ~10 minutes

Possibly related issues

• [ponytail#214] Windows stdin timeout — hooks must never block on Windows #71: Addresses making stdin reads in hook paths timeout-safe, matching this PR's core change.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title clearly matches the main hook changes: stdin timeout, stderr logging, and /agentflare reporting.
Description check	✅ Passed	Covers the main changes, tickets, and test result; only the template headings and reviewer notes are missing.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
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.

✨ Finishing Touches

📝 Generate docstrings

• Create stacked PR
• Commit on current branch

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch fix/windows-stdin-timeout

---

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

[coderabbitai]

🧹 Nitpick comments (2)

src/hook.rs (2)

97-103: 📐 Maintainability & Code Quality | 🔵 Trivial | ⚡ Quick win

Duplicated timeout-handling boilerplate across both call sites.

The match read_stdin_timeout(1000) { Some(s) if !s.is_empty() => s, _ => { eprintln!(...); return; } } block is repeated verbatim (only the log label differs) in pre_tool_use and prompt_submit, and the 1000 ms timeout is a duplicated magic literal. Consider extracting a small helper to reduce duplication and centralize the timeout constant.

♻️ Suggested refactor

+const STDIN_TIMEOUT_MS: u64 = 1000;
+
+fn read_stdin_or_skip(label: &str) -> Option<String> {
+    match read_stdin_timeout(STDIN_TIMEOUT_MS) {
+        Some(s) if !s.is_empty() => Some(s),
+        _ => {
+            eprintln!("[agentflare] {label}: stdin timeout or empty — skipping");
+            None
+        }
+    }
+}

Then in pre_tool_use:

-    let input = match read_stdin_timeout(1000) {
-        Some(s) if !s.is_empty() => s,
-        _ => {
-            eprintln!("[agentflare] PreToolUse: stdin timeout or empty — skipping");
-            return;
-        }
-    };
+    let Some(input) = read_stdin_or_skip("PreToolUse") else { return };

And similarly in prompt_submit with the "UserPromptSubmit" label.

Also applies to: 155-161

🤖 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 `@src/hook.rs` around lines 97 - 103, Extract the repeated stdin timeout
handling in pre_tool_use and prompt_submit into a small shared helper that wraps
read_stdin_timeout and the empty-input check, so both call sites just pass their
log label. Replace the duplicated 1000 ms magic literal with a central timeout
constant used by that helper, and keep the existing eprintln!/return behavior
for the "PreToolUse" and "UserPromptSubmit" paths.

---

269-273: 📐 Maintainability & Code Quality | 🔵 Trivial | 💤 Low value

Test doesn't verify meaningful behavior.

read_stdin_timeout_does_not_panic only checks the call doesn't panic; it exercises neither the None (timeout) branch nor a populated-input branch. Given this is the core fix for tickets #71/#70, consider parameterizing the read logic (e.g., accept a generic Read or an injected timeout duration for testing) so the timeout path can be asserted deterministically instead of relying on real stdin timing in tests.

🤖 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 `@src/hook.rs` around lines 269 - 273, The read_stdin_timeout test currently
only proves the call does not panic and does not validate either the timeout
path or the successful input path. Refactor the stdin-reading logic in
read_stdin_timeout to be testable with an injected reader or configurable
timeout so you can deterministically assert the None branch and a
populated-input branch without relying on real stdin timing. Update
read_stdin_timeout and its tests, including read_stdin_timeout_does_not_panic,
to cover the actual behavior of the timeout handling.

🤖 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.

Nitpick comments:
In `@src/hook.rs`:
- Around line 97-103: Extract the repeated stdin timeout handling in
pre_tool_use and prompt_submit into a small shared helper that wraps
read_stdin_timeout and the empty-input check, so both call sites just pass their
log label. Replace the duplicated 1000 ms magic literal with a central timeout
constant used by that helper, and keep the existing eprintln!/return behavior
for the "PreToolUse" and "UserPromptSubmit" paths.
- Around line 269-273: The read_stdin_timeout test currently only proves the
call does not panic and does not validate either the timeout path or the
successful input path. Refactor the stdin-reading logic in read_stdin_timeout to
be testable with an injected reader or configurable timeout so you can
deterministically assert the None branch and a populated-input branch without
relying on real stdin timing. Update read_stdin_timeout and its tests, including
read_stdin_timeout_does_not_panic, to cover the actual behavior of the timeout
handling.

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro Plus

Run ID: 5666e382-0309-421e-adc5-5aaff5dc3eac

📥 Commits

Reviewing files that changed from the base of the PR and between 7af46db and a8518f1.

📒 Files selected for processing (1)

• src/hook.rs