Verify

Verify

Ask Claude Code to check whether previously reported audit findings have been resolved. Claude re-reads the files at the reported locations and renders a verdict per issue. Codex provides the issue list; Claude provides the verdict.

When to Use

• After Codex has applied fixes from a $audit run
• To confirm a manual fix was applied at the correct location
• As the final step of a $audit-fix cycle

Arguments

Call Pattern

Option A: Continue the audit session (preferred)

If {audit_session_id} is available from a prior $audit or $audit-fix run:

Copy code
mcp__claude-code__claude_code_reply:
  session_id: {audit_session_id}
  prompt: |
    The following issues from your audit have been addressed. Verify each one.

    ISSUES TO VERIFY:
    {paste findings in "file:line | severity | issue description" format}

    For each issue report one of:
    - FIXED — issue resolved fully, no new problems introduced
    - NOT FIXED — issue still present at the reported location (explain why)
    - PARTIAL — partially addressed; describe what remains
    - REGRESSED — fix introduced a new problem (describe it)

    Read the files at the reported locations. Do not assume a fix is correct
    without reading the current code.

Option B: Fresh verification session

When no prior session is available:

Copy code
mcp__claude-code__claude_code:
  prompt: |
    Verify whether the following issues have been fixed.

    ISSUES TO VERIFY:
    {paste findings in "file:line | severity | issue description" format}

    For each issue:
    1. Read the file at the reported location
    2. Check if the issue is resolved
    3. Report: FIXED / NOT FIXED / PARTIAL / REGRESSED

    For NOT FIXED and PARTIAL: explain what is still wrong.
    For REGRESSED: describe the new problem introduced.

    Apply independent judgment — do not accept surface-level changes that
    mask the underlying issue.
  cwd: {project working directory}
  effort: high
  permissionMode: plan

Save the returned session_id as {verify_session_id}.

Output Format

Result: N fixed, N not fixed, N partial, N regressed

Failure Handling

A Claude-backed call that does not return a usable result is terminal for this skill. Stop and report; do not improvise a way to keep going.

This covers every non-success outcome: a runtime or transport error, a timeout, a rate limit, an empty or truncated result, and a result that does not match the shape this skill asked for. Absence of an error is not success — a call that returns nothing has failed, however it exited.

Never report a failed call as a result. It is not CLEAN, not "no findings", not a completed verification, not an empty plan. A verification that did not run is UNVERIFIED, and saying so is the entire value of the step.

Do not recover on your own initiative. Unless the user asks for it: do not retry the call, do not open a replacement session, do not switch to another Claude execution path (a different MCP tool, a different skill, a direct invocation), and do not fall back to doing the work yourself in place of the delegate. A failed claude_code_reply in particular must not silently become a new claude_code session — that discards the conversation the reply was meant to continue, and the original work may still be running.

Preserve the session. Report any session_id the call returned, even on failure, so the user can resume that exact session rather than start over.

Say what may still be running. A timeout or transport error does not establish that the work on the other side stopped. Say so explicitly, so the user does not start a second attempt that overlaps the first.

Then stop according to what the call could have changed:

Notes

• Always prefer Option A when a session is available — Claude retains full audit context and can make more precise verdicts
• If all issues are FIXED, report success and suggest committing
• If any are NOT FIXED or REGRESSED, feed those back into $audit-fix or fix manually

Example Invocations