All skills
simota avatar

/sigil

@35ffd55
by shingo imotasimota/agent-skills85 stars
15

Designing a repository's project-local operating layer and generating its skills, recipes, workflows, and routing map. Not for global ecosystem agents (Architect) or runtime execution (Nexus).

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

This session only. Nothing lands on disk.

referenceoperating-layer-handoffs.md

≈1.2k tokens on demand. Your agent reads this file only when SKILL.md points to it.

Delegation Handoffs

Purpose: Per-component handoff specs and the project routing-map format. Sigil[blueprint] designs; these specs hand each piece to its owner. Read when: You are in DELEGATE, building the routing map (map mode), or finalizing a blueprint.

Contents

  • Principle
  • Routing map format
  • Handoff payloads by owner
  • Escalation to Architect
  • Completeness rule

Principle

Sigil[blueprint]'s output is a design plus a set of handoffs — never an installed artifact. Every component of the blueprint (each skill, recipe, workflow, hook, placement decision) must name exactly one owning agent and the payload that agent needs to execute. A component with no owner is an incomplete design and fails VERIFY.

Use the canonical envelope in _common/HANDOFF.md; the fields below are the Sigil[blueprint]-specific payload each owner receives.


Routing map format

The routing map answers "which agent or skill owns which repo task domain" and is consumed by Nexus. Every task domain has exactly one primary owner. Ecosystem-owned domains point at the existing agent (defer, never duplicate).

routing_map:
  # task domain        : owner (project skill or ecosystem agent)
  "db schema design"   : schema            # ecosystem
  "db query tuning"    : tuner             # ecosystem
  "trpc procedure"     : new-trpc-procedure  # project-local (Sigil authors)
  "app route"          : new-app-route       # project-local
  "security review"    : judge             # ecosystem — deferred
  "ship a feature"     : recipe:ship-feature
  "cut a release"      : workflow:release-train

Conflict rule: if two owners claim one domain, narrow scope until each domain is singly owned, or merge. Surface unresolved conflicts as VERIFY failures.


Handoff payloads by owner

Sigil[blueprint] → Sigil (SIGIL_BLUEPRINT_TO_SIGIL_HANDOFF) — author skill bodies

skill_specs:
  - name: new-trpc-procedure
    responsibility: "Generate a tRPC procedure with Zod input + colocated test"
    conventions: [import-alias "@/", zod-in src/schemas, vitest]
    type_hint: Micro | Full
    description_requirements:
      - third person, gerund-form name
      - include negative triggers ("Don't use when…")
    install_target: project (.claude/skills + .agents/skills)

Sigil[blueprint] states what the skill must do and which conventions to mirror; Sigil decides Micro/Full and writes the body, scores it, and installs.

Sigil[blueprint] → Nexus (SIGIL_BLUEPRINT_TO_NEXUS_HANDOFF) — register routing + chains

nexus_registration:
  routing_map: <the map above>
  recipes:   [<recipe definitions from recipe-design.md>]
  workflows: [<workflow definitions with topology + phases>]

Sigil[blueprint] → Orbit (SIGIL_BLUEPRINT_TO_ORBIT_HANDOFF) — autonomous loop specs

loop_spec:
  workflow: release-nightly
  goal: "nightly release if main is green"
  stop_condition: "no merged PRs since last run"
  operation_contract: <inputs, success check, recovery>

Sigil[blueprint] → Hone (SIGIL_BLUEPRINT_TO_HONE_HANDOFF) — enforcement hooks

hook_specs:
  - event: PreToolUse
    matcher: "git commit"
    enforce: "block direct commits to main"
    rationale: "every-time guard — not a skill"

Sigil[blueprint] → Grove (SIGIL_BLUEPRINT_TO_GROVE_HANDOFF) — placement

placement:
  recipes_doc: docs/recipes/ or CLAUDE.md section
  monorepo_scope: per-package paths if applicable
  # skill install dirs (.claude/skills/ + .agents/skills/) are owned by Sigil, not Grove —
  # never route them here; doing so creates an L6 double-owner on the install path

Escalation to Architect

When SURVEY/MAP surfaces a gap that is not project-specific — a capability that would help many repositories — escalate to Architect for a global ecosystem agent instead of designing a project-local one. Require evidence it generalizes (the same need would recur across unrelated repos); otherwise keep it project-local.

SIGIL_BLUEPRINT_TO_ARCHITECT_HANDOFF:
  gap: "<capability>"
  why_global: "<evidence it generalizes beyond this repo>"
  project_workaround: "<what the project-local layer does meanwhile>"

Completeness rule

Before emitting _STEP_COMPLETE, confirm:

  • every skill in the suite plan → a Sigil spec,
  • every recipe and workflow → a Nexus registration entry,
  • every enforcement point → a Hone hook spec,
  • every loop → an Orbit loop spec,
  • every task domain → exactly one routing-map owner.

Any unowned component blocks delivery.

Source: SKILL.md on GitHub

1 warning13d5 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    The skill is a project-local development tool designed to analyze codebase context and generate AI agent skills. It uses shell commands and local file access to detect tech stacks and conventions. While it includes safety guardrails against prompt injection and credential exposure, it possesses an attack surface for indirect prompt injection, where a malicious repository could influence the generation of executable skill instructions or configurations.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    2/12 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at 35ffd55. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 3 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/sigil