agentleFS
solo-istsolo-ist/prose/.agents/skills/pipeline-orchestrate/SKILL.md

pipeline-orchestrate

Skill8 starsChanged 5 months agoMITView on GitHub

Merge scorer and PE signals to route PRs. Applies routing matrix (auto-fix / auto-fix-verify / hitl-light / hitl-full) based on score thresholds and risk levels.

The file

We split this skill into the rules it gives an agent

3 rules. About 0 are common, 1 is distinctive, and 2 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈1

    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:

    Read scorer and PE comments from the PR

    line 50
  2. Very specific:

    If verdict is auto-fix or auto-fix-verify, dispatch pipeline-fix.yml

    line 55
  3. Very specific:

    If the same PR also touched src/main/ipc.ts:

    line 66

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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

Pipeline Orchestrator

Merges scorer and PE analysis signals to determine the final routing verdict for a PR.

Usage

/pipeline-orchestrate <pr-number>

Routing Matrix

Source of truth: run-orchestrator.mjs route() function.

Score (1-10)PE RiskRouteAction
1-3LOWauto-fixCodex pushes fix commit
1-3MEDIUM+hitl-lightPost analyses, label needs-review
4-6LOWauto-fix-verifyFix + request human verification
4-6MEDIUM+hitl-fullFull analyses, assign human
7+Anyhitl-fullLabel complex, full human review
AnyCRITICALhitl-fullAlways escalate
Anyprivilegedhitl-fullSecurity hard gate — always escalate

Label Mapping

VerdictLabels Applied
auto-fixauto-fix-queued
auto-fix-verifyauto-fix-queued, needs-review
hitl-lightneeds-review
hitl-fullcomplex, needs-review

Security Hard Gate

PRs touching privilege-boundary paths auto-route to hitl-full regardless of score:

  • src/main/** — full system access
  • src/preload/** — context bridge / attack surface
  • electron-builder.*, electron.vite.config.* — build/packaging config

With the security-gate short-circuit (383-2), these PRs skip scorer+PE entirely — the pipeline-triage.yml security-gate-check job detects the <!-- security-gate: true --> sentinel in the review and posts hitl-full directly.

Workflow

  1. Read scorer and PE comments from the PR
  2. Parse sentinel JSON from both (<!-- scorer-output: {...} --> and <!-- pe-output: {...} -->)
  3. Apply routing matrix
  4. Post verdict comment with <!-- orchestrator-verdict: auto-fix|auto-fix-verify|hitl-light|hitl-full -->
  5. Apply GitHub labels per the label mapping table
  6. If verdict is auto-fix or auto-fix-verify, dispatch pipeline-fix.yml

Decision Trace Example

Given a PR that touches src/renderer/components/editor/Editor.tsx only:

  • Scorer returns: {"score": 4, "threshold": "review"}
  • PE returns: {"risk": "LOW", "privileged": false, "concerns": 1}
  • Path: score 4 >= 4, risk LOW → auto-fix-verify
  • Labels: auto-fix-queued, needs-review
  • Pipeline dispatches pipeline-fix.yml, then requests human verification

If the same PR also touched src/main/ipc.ts:

  • PE returns: {"risk": "MEDIUM", "privileged": true, "concerns": 2}
  • Path: privileged = true → immediate hitl-full (score doesn't matter)
  • Labels: complex, needs-review
  • No auto-fix dispatched

Output Format

Machine-readable sentinel: <!-- orchestrator-verdict: auto-fix|auto-fix-verify|hitl-light|hitl-full -->

Human-readable summary with routing rationale referencing scorer dimensions and PE risk assessment.

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.