skill-debug

1> **Host: Codex CLI** — This skill was designed for Claude Code and adapted for Codex.
2> Cross-reference commands use installed skill names in Codex rather than `/octo:*` slash commands.
3> Use the active Codex shell and subagent tools. Do not claim a provider, model, or host subagent is available until the current session exposes it.
4> For host tool equivalents, see `skills/blocks/codex-host-adapter.md`.
5 
6 
7# Debugging
8 
9Read `skills/blocks/engineering-method-selection.md` from the installed plugin
10for review admission. Natural-language requests and `--peer-review` share that
11policy. Honor host-only requests; risk alone does not authorize paid usage.
12 
13Run the investigation on the current host. Routine debugging makes zero
14additional provider dispatches. Use a bounded external reviewer only for
15`--peer-review`, an explicit independent-review request, or an existing risk
16policy.
17 
18<HARD-GATE>
19DO NOT CHANGE PRODUCTION BEHAVIOR BEFORE REPRODUCING THE SYMPTOM AND TESTING A
20NAMED ROOT-CAUSE HYPOTHESIS.
21</HARD-GATE>
22 
23Read and apply `skills/blocks/debug-feedback-loop.md` from the installed plugin
24root.
25 
26Start with the user's observable symptom. Reproduce it, retain its failure
27signature while minimizing the scenario, test one named hypothesis at a time,
28and verify both the minimal reproduction and the original scenario after the
29fix. Do not treat a nearby passing helper test as proof.
30 
31For a race, use a synchronization barrier and a fixed run or time budget. For an
32…