| name | trellis-skill-assimilator |
| description | Extract reusable engineering capabilities from external GitHub skill projects and produce safe Trellis absorption plans. Use when the user provides a GitHub repository URL, skill project URL, open-source agent skill collection, or asks to compare or update projects such as superpowers, gsd, CodeStable, loop-engineering, ponytail, or similar. Analyze README/SKILL.md/references/scripts/license and produce source analysis, capability extraction, absorption card, Trellis target mapping, risk judgment, update-scan delta report, and verification scenario recommendations. Do not copy third-party skill text directly; absorb capability patterns, workflow structure, and verifiable design only.
|
Trellis Skill Assimilator
Overview
Convert useful patterns from external GitHub skill projects into governed Trellis capabilities. The goal is not to copy a skill. The goal is to answer:
- What engineering problem does the project solve?
- Which Trellis capabilities are missing or weak?
- Which capabilities should be absorbed, watched, or rejected?
- Which Trellis stage, reference, or gate should receive the capability?
- How will we verify that the absorption improves output quality?
Guardrails
- Do not copy large passages from third-party skills.
- Absorb by capability, not by project name.
- Do not absorb a capability without a Trellis target.
- Do not merge a capability without a verification scenario.
- When license is unclear, strong copyleft, commercially restricted, or untrusted, only summarize ideas; do not move text, scripts, or templates.
- Do not directly modify Trellis core skills; first produce an absorption plan and verification recommendation, then wait for user confirmation.
- Do not put absorption governance into
trellis-zero-to-mvp, trellis-mvp-to-delivery, or trellis-delivery-controller.
Modes
Initial Assimilation
Use for first-time analysis of an external GitHub skill project.
- Read
references/source-analysis-template.md.
- Analyze repository README, skill directories,
SKILL.md, references, scripts, docs, and license.
- Read
references/capability-taxonomy.md and classify findings as capabilities.
- Read
references/trellis-target-map.md and map capabilities to Trellis targets.
- Read
references/license-safety-policy.md and decide the absorption boundary.
- Use
references/absorption-card-template.md to output the absorption card.
Update Scan
Use when a previously analyzed project changes.
- Read the existing absorption card or user-provided
last_analyzed_ref.
- Compare current ref against
last_analyzed_ref for README, SKILL.md, references, scripts, docs, and license changes.
- Read
references/update-scan-policy.md and analyze only new capabilities, changed capabilities, risk changes, or updates to already absorbed capabilities.
- Output a delta absorption report.
Output
Every run must output:
- Source Analysis.
- Capability Extraction.
- Absorption Card.
- Trellis Target Map.
- License and Safety Notes.
- Evaluation Requirements.
- Recommendation:
absorb, watch, reject, or needs-human-review.
If recommending absorption, also provide:
- Trellis files to add or modify.
- Purpose of each change.
- New verification scenario.
- Rollback condition.
- Whether strong-model or human review is required.
References
references/source-analysis-template.md - Read when analyzing external repository structure and capability.references/absorption-card-template.md - Read when producing an absorption card.references/capability-taxonomy.md - Read when classifying and naming capabilities.references/trellis-target-map.md - Read when mapping to Trellis targets.references/license-safety-policy.md - Read when deciding license and content movement boundaries.references/update-scan-policy.md - Read when scanning an already analyzed project for changes.
