agentsclimarketplace

Reviewer protocol

Skill gustavo-meilus/superpipelines/.claude/skills/superpipelines/parity-test-b/reviewer-protocol

Loaded by the reviewer agent of parity-test-b to supply its operating protocol. Not user-invocable.From its SKILL.md

Install
npx -y skills add gustavo-meilus/superpipelines --skill reviewer-protocol

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 4 stars4 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

SKILL.md

2.6 KB, 648 tokens by cl100k_base, as published. Nobody here has run it

Reviewer — Protocol

Step 2 of 3 in the parity-test-b pipeline. Reads findings.json from the temp directory, validates schema completeness and correctness. READ-ONLY — this agent runs with permissionMode: plan and disallowedTools: Write, Edit, Bash. Verdict is communicated exclusively via terminal output text.

Inputs (from dispatch context)

  • findings_path — absolute path to findings.json (written by the analyzer step).
  • state_path — absolute path to pipeline-state.json (read-only reference — DO NOT write).
  • run_id — current run ID string.
  • root — resolved scope root ({ROOT}).

Protocol

PHASE 1: VALIDATE FINDINGS FILE

  1. Confirm findings_path exists. If missing: emit BLOCKEDfindings.json not found. Analyzer step may not have completed. — STOP.
  2. Parse as JSON. If invalid: emit BLOCKED — STOP.
  3. Validate required fields: source_path (string), analyzed_at (string), total_keys (number ≥ 0), issue_count (number ≥ 0), keys (array, length == total_keys). If any missing or wrong type: emit BLOCKED — STOP.

PHASE 2: VALIDATE KEY ENTRIES

For each entry in keys:

  1. Confirm presence of: key (string), type (string), is_null (boolean), type_inconsistent (boolean), issue_count (number ≥ 0).
  2. Confirm type is one of: string, number, boolean, object, array, null.
  3. Accumulate any validation failures.

PHASE 3: RENDER VERDICT

Determine verdict:

  • approved — zero failures.
  • approved_with_concerns — schema valid but concerns (e.g., all keys are null, zero total_keys).
  • rejected — one or more structural failures.

Render verdict block in terminal output:

REVIEWER VERDICT: {verdict}
COMPLETENESS: {short summary}
CONCERNS: {list or "none"}
NOTES: {brief notes}

PHASE 4: EMIT STATUS

  • DONE — verdict: approved.
  • DONE_WITH_CONCERNS — verdict: approved_with_concerns.
  • BLOCKED — verdict: rejected or validation failure.

Invariants

  • NEVER write to any file — permissionMode: plan + disallowedTools: Write, Edit, Bash enforce structural write-deny.
  • NEVER use hardcoded platform paths.
  • ALWAYS render the REVIEWER VERDICT block before emitting terminal status.
  • ALWAYS validate every key entry — do not short-circuit on first failure.
  • Emit exactly one terminal status: DONE / DONE_WITH_CONCERNS / BLOCKED.

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 326,569. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.