All skills
simota avatar

/architect

@965f4f9
by shingo imotasimota/agent-skills85 stars
15

Designing new skill agents via gap analysis, overlap detection, SKILL.md + reference generation, and Nexus integration. Not for task orchestration (Nexus) or format-only audits (Gauge).

Use this Skill: https://skilld.dev/gh/simota/agent-skills/architect

This session only. Nothing lands on disk.

referenceoverlap-detection.md

≈566 tokens on demand. Your agent reads this file only when SKILL.md points to it.

Functional Overlap Detection

Read during ANALYZE. Compare the proposed primary action, inputs and outputs with actual current SKILL.md definitions. Use _common/BOUNDARIES.md for authoritative ownership; do not substitute a copied roster or name similarity for a capability comparison.

Matching and Score

Capability relationship Match value
Exact 1.0
Synonym / equivalent responsibility 0.8
Partial 0.5
Related but different responsibility 0.2
None 0.0

Overlap (%) = sum(capability match values) / total capabilities × 100.

Record the capability inventory, each pairing and the denominator used. The original rubric does not specify normalization for unequal inventories: expose that ambiguity and review the ownership evidence rather than presenting an arbitrary normalization as canonical. Do not reuse example percentages as observations.

Assess three dimensions: functional (actions/inputs/outputs), domain (technical/business/code scope), and collaboration (chain position/partners/blocking dependencies). Shared domain alone is not shared ownership.

Decision Bands

Score Decision
0–10% Proceed
10–20% Note overlap
20–30% Review differentiation
≥30% and <50% Ask first; document differentiation
≥50% Reject by default; propose the existing-owner alternative

At a boundary, the higher approval requirement applies. SKILL.md owns the approval/rejection gate; scoring uncertainty near a gate must not be used to bypass it.

Report Contract

Report the proposed name, purpose and category; the top three existing overlaps; matched and unique capabilities; per-agent scores with their evidence/denominator; the differentiating outcome; and a recommendation with alternatives. An overlap table without a concrete remaining gap is not a creation justification.

Exceptions

A deliberate replacement, domain-specific variant or explicit request for overlapping agents requires user confirmation and a written reason. Record the overlap percentage, existing owner and distinguishing scope. A replacement also needs a migration plan; it does not authorize silent deletion or immediate rerouting.

Source: SKILL.md on GitHub

2 warnings13d5 checks · Risk MEDIUM
  • Gen Agent Trust Hub13d

    The skill features a self-evolution capability that allows it to autonomously modify its own reference files and instructions. It also acts as an agent design tool, which introduces a surface for indirect prompt injection based on untrusted user requirements.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    5/18 files flagged

  • ZeroLeaks5mo

    1 finding · Score: 69/100

Signed by skilld at 965f4f9. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/architect