straymark-audit-prompt

name	straymark-audit-prompt
description	Generate the unified audit prompt for a Charter at the canonical filesystem location. The operator then opens N auditor-side CLIs (gemini-cli, claude-cli, copilot-cli, etc.) and invokes /straymark-audit-execute in each — no copy/paste. Counterpart of /straymark-audit-review.

StrayMark Audit Prompt Skill

Generate the unified audit prompt for a Charter and write it to the canonical filesystem location. The operator then opens auditor-side CLIs in the same repo and invokes /straymark-audit-execute — the audit prompt is read directly from disk by each auditor; the operator never copies/pastes.

When to invoke

Use this skill when the developer agreed to run an external audit at the Charter checkpoint (see .straymark/00-governance/AGENT-RULES.md § Audit checkpoint).

The Charter should be in in-progress or declared status — auditing closed Charters is allowed but atypical (warn the operator and proceed only on confirmation).

Audit a stable state. Generate the prompt only when the state under audit is stable: the PR's CI is green, the working tree is clean, and commits are pushed. --prepare embeds the git diff at generation time, so if in-flight CI later forces a fix (or uncommitted work changes the tree), the prompt goes stale and the auditors review a revision that no longer exists. Sequence: PR → CI green → --prepare → launch auditors. The CLI warns (does not block) when the tree is dirty or has unpushed commits.

Instructions

1. Resolve the Charter

Argument: a Charter identifier (CHARTER-04, 04, or full id with slug).

straymark charter status <CHARTER-ID>

Verify the Charter exists and capture its status. If status: closed, surface a one-line warning to the operator and ask whether to proceed.

2. Generate the unified audit prompt

straymark charter audit <CHARTER-ID> --prepare

The CLI writes the resolved prompt to:

.straymark/audits/<CHARTER-ID>/audit-prompt.md

Multi-round audits (--round). For a multi-phase Charter audited in more than one round (e.g. one external-audit round per code-heavy phase), pass --round <label> so the whole triad (prompt, reports, review) is namespaced under .straymark/audits/<CHARTER-ID>/<label>/ instead of the flat charter root. This keeps rounds from overwriting each other and scopes the report-*.md glob to just this round. Omit --round for single-round Charters. Thread the same --round <label> through /straymark-audit-execute and /straymark-audit-review. See .straymark/00-governance/AUDIT-ROUNDS-PATTERN.md.

The prompt is self-contained: it embeds the Charter content, the originating AILOGs, the git diff over the resolved range (default origin/main..HEAD, falls back to HEAD~1..HEAD if no upstream is reachable), and the discipline rules (REGLA ABSOLUTA — SOLO LECTURA, evidence-citation, severity calibration). The prompt template lifts the seven universal sections from Sentinel's pre-StrayMark audit skill and parameterizes the project-specific hardcodes.

Multi-batch Charters — pass an explicit --range. When auditing one phase of a Charter whose earlier phases already merged to the base branch, the default origin/main..HEAD excludes the already-merged commits and the prompt silently under-covers the phase. Pass --range <charter-first-commit>..HEAD so all of the phase's commits are in the diff. The CLI prints a warning when it detects completed batches in the Charter's Batch Ledger and no explicit range was given.

The CLI does NOT invoke any LLM. It only resolves placeholders.

3. Notify the operator

After the CLI succeeds, print this guidance verbatim (substituting <CHARTER-ID>):

Audit prompt prepared for <CHARTER-ID>.

  Location: .straymark/audits/<CHARTER-ID>/audit-prompt.md
  (informational — you don't need to copy this path anywhere)

Next steps:

  1. Open one or more auditor-side CLIs (gemini-cli, claude-cli,
     copilot-cli, codex-cli — whatever you have) in this repo. Each
     CLI session uses its own model; recommendation is at least 2
     auditors of DIFFERENT model families, so cross-family blind
     spots become signal when their findings converge.

  2. In each auditor CLI, invoke:

         /straymark-audit-execute <CHARTER-ID>

     The skill is already installed (straymark init copies it to all
     three platform locations). It reads the prompt from disk
     automatically — you do not need to copy/paste anything.

  3. When and only when ALL audits you commissioned have completed,
     return to this main agent and run:

         /straymark-audit-review <CHARTER-ID>

     Reviewing with incomplete reports gives you a partial consolidated
     analysis that you will have to discard or re-run.

Notes

• This skill is orchestration-only. It does NOT invoke LLM APIs, decide which models the operator uses, or wait for responses. It runs the CLI's --prepare, points the operator at the next-step skill, and exits.
• Re-running the skill on the same Charter regenerates the prompt at the same path. It does NOT touch operator-saved reports under .straymark/audits/<CHARTER-ID>/report-*.md.
• Heterogeneity inter-family is recommended but not enforced in v0; the operator decides the model pairing across the N CLIs they open.
• For the rationale on cross-family auditors, see dist/.straymark/audit-prompts/audit-prompt.md § Tu rol or Propuesta/straymark-cli-roadmap.md §5.2 in the upstream repo.