agentleFS
kvantatechkvantatech/RedLine/.github/skills/triage-perf-verdict/SKILL.md

triage-perf-verdict

Skill1 starsChanged 3 months agoMITView on GitHub

Type: skill (1 model narration) · Used by: perf-run-one · Status: LIVE You will draft a Jira regression ticket for a red performance verdict. Follow these steps exactly. Inputs (passed by the workflow): - run_id — the run identifier - grafana_url — deep-link URL from O7 link-grafana-panel; empty string "" if O7 was skipped or OBSERVE!=1 Read both files: - reports/<run_id>/verdict.json - reports/<run_id>/contract.json From the verdict and contract, write exactly one sentence describing what regressed and by how much. Rules:…

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 2 are distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If multiple endpoints are red, name all of them in the same sentence.

    line 29
  2. Very specific:

    All fields except [NARRATION] are deterministic — copy the values verbatim, no invention.

    line 34
  3. Very specific:

    Write the assembled draft to reports/<run_id>/jira-draft.md.

    line 72
  4. Distinctive:

    Read and Write only — one model narration call, no MCP.

    line 86

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

triage-perf-verdict

Draft Jira ticket text for a red verdict — 1 AI narration line + deterministic template.

Type: skill (1 model narration) · Used by: perf-run-one · Status: LIVE


Prompt

You will draft a Jira regression ticket for a red performance verdict. Follow these steps exactly.

1 — Read inputs

Inputs (passed by the workflow):

  • run_id — the run identifier
  • grafana_url — deep-link URL from O7 link-grafana-panel; empty string "" if O7 was skipped or OBSERVE!=1

Read both files:

  • reports/<run_id>/verdict.json
  • reports/<run_id>/contract.json

2 — Write ONE narration line (model call)

From the verdict and contract, write exactly one sentence describing what regressed and by how much. Rules:

  • Neutral tone. No blame, no guessing at root cause.
  • Name the endpoint, the p95, the threshold, and the delta.
  • Example: "GET /api/config p95 climbed to 1250ms (+350ms, +38.9% over the 900ms red threshold) on demo-web STG as of 2026-06-09."
  • If multiple endpoints are red, name all of them in the same sentence.

3 — Assemble the draft ticket

Fill the template below with values from the verdict/contract + the narration line. All fields except [NARRATION] are deterministic — copy the values verbatim, no invention.

h2. Performance Regression — Auto-detected

Team: <team>
Env: <env>
Run ID: <run_id>
Verdict: red
Date: <started_at from contract, date part only>
Summary: [NARRATION]

h3. Failing endpoints

| Endpoint | p95 ms (Result) | Red threshold ms | Delta |
| <name> | <p95_ms> | <p95_red_ms> | +<ms> (+<pct>%) |

h3. Checks

- Total iterations: <contract.iterations>
- Check pass rate: <contract.checks.pass_rate × 100>%
- Error rate: <endpoint error_rate>

h3. Links

- Grafana: <grafana_url if non-empty, else "[unavailable]">
- Run ledger: state/run-ledger.jsonl (filter run_id=<run_id>)

h3. Action required

Human review required before this ticket is filed. Route to reviewer subagent next.

Template is Jira wiki markup (h2./h3., pipe-delimited tables) — NOT markdown. This is the format that file-perf-regression-jira copies verbatim into the Jira description field.

4 — Write output

Write the assembled draft to reports/<run_id>/jira-draft.md.

Return:

{
  "draft_path": "reports/<run_id>/jira-draft.md",
  "narration": "<the one-sentence narration>"
}

Tools

Read and Write only — one model narration call, no MCP.


Data

  • Inputs: run_id (string), grafana_url (string — from O7; empty string if unavailable)
  • Reads: reports/<run_id>/verdict.json, reports/<run_id>/contract.json
  • Writes: reports/<run_id>/jira-draft.md

Hard rules

  1. Exactly one model narration line — the rest of the ticket is templated from the data.
  2. Draft only — file-perf-regression-jira is the separate human-gated filing step.
  3. Only runs on red verdicts — green and no-baseline never reach this skill.
  4. p95 everywhere — never quote p50 or p99 in the narration.

Frozen skill registry: AGENTS.md in this repo

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.