run legion
Official@runlegion
Run Legion provides quality-gated pull request review pipelines and double-diamond service design document pipelines for repository engineering teams.
Agent Skills by run legion
Showing 14 vetted skills indexed across 1 GitHub repositories.
runlegionlegion-review
Reviews open PRs against issue specs using parallel dimension agents and adversarial finding verification.
runlegionsd-write-blueprint
Writes schema-valid service blueprint documents mapping frontstage and backstage layers to journey steps.
runlegionlegion-distill
Validates each block of durable prose for clarity, self-containment, and consistency before storage.
runlegionlegion-simplify
Reviews branch diffs for structural code quality issues and records a validated per-file quality gate.
runlegionsd-discover
Validates service-design claims against real discourse evidence and produces a schema-valid Discovery document.
runlegionsd-write-journey
Writes schema-valid service-design journey documents with evidence-traced emotional curves.
runlegionsd-write-spec
Writes IEEE 830-style FR and NFR requirement documents from intent and discovery inputs.
runlegionsd-write-persona
Writes one schema-valid service-design persona traced to Discovery insights and intent fields.
runlegionlegion-memory
Queries Legion memory reflections before searching the codebase for answers.
runlegionsd-service-design
Orchestrates a service design pipeline from intent review through discovery to personas, journeys, and blueprints.
runlegionlegion-verify
Validates issue acceptance criteria against diffs and test output with cited evidence verdicts.
runlegionsd-ecosystem-imagine
Generates a schema-validated ecosystem document mapping actors, channels, and value exchanges for a service intent.
runlegionlegion-pr-write
Compose and validate PR bodies mapping issue acceptance criteria to code changes.
runlegionsd-intent-review
Derives a testable research agenda of services and claims from a repository intent document.
Frequently Asked Questions About run legion
FAQPage SchemaWhat tasks can I perform with run legion's skills?▼
You can run a gated PR pipeline (simplify, pr-write, review, verify) that checks branch diffs for structural issues, maps acceptance criteria to changes, adversarially verifies findings, and blocks completion on failed evidence. Separately, you can run a service design pipeline producing intent, discovery, ecosystem, persona, journey, blueprint, and IEEE 830-style FR/NFR spec documents.
Who are run legion's skills designed for?▼
They target solo engineers and small engineering teams lacking dedicated QA, plus agents conducting repository service design. The verify skill explicitly positions itself as the QA a solo operator does not have, while the sd-* skills serve operators defining services from research evidence before writing code.
How do the legion quality gates work in practice?▼
Each stage records a gate via commands like `legion quality-gate record` and `legion pr write-check`; downstream steps refuse to proceed without a clean prior gate. Simplify runs on every feature branch before PR creation, review runs after `legion pr create`, and verify runs last, hard-blocking on any failed acceptance criterion.
What are the prerequisites and dependencies for these skills?▼
Skills require the legion CLI and its document store, with allowed tools limited to Bash, Read, Write, and Task. Service design steps must run in order: intent-review before discover, discovery before persona and ecosystem, journey before blueprint. The sd-write-spec skill has its own entry point outside the sd-service-design pipeline.
How does the service design pipeline handle missing evidence?▼
It never guesses: sd-discover parks on a missing corpus or open question, contradicted claims become findings the operator rules on rather than being cut, and steps that cannot finish park and resume. Every persona, journey, and blueprint statement must trace to a Discovery insight or intent field.