Sigil Reference Index
Every reference/ file sigil owns, and the condition that makes it worth
reading. sigil/SKILL.md keeps only the shared-contract rows and a pointer here.
Read this when you need a reference and the Recipe registry did not already name it, or when scanning what this skill can consult at all.
| Reference | Read this when |
|---|---|
reference/context-analysis.md |
SCAN on any project or refresh to detect stack, conventions, monorepo layout, existing skills, and sync drift. |
reference/skill-catalog.md |
Ranking candidates in DISCOVER to map frameworks to likely high-value skills and migration paths. |
reference/skill-templates.md |
Drafting any new skill in CRAFT to choose Micro vs Full, apply templates, and preserve required structure. |
reference/validation-rules.md |
Scoring before install or after updates to apply structural checks, rubric scoring, and validation reporting. |
reference/evolution-patterns.md |
Updating stale skills to choose lifecycle state, trigger handling, and update strategy. |
reference/advanced-patterns.md |
Handling variants, monorepos, or composed skills with conditional branches, variable substitution, scoping, and composition rules. |
reference/skill-effectiveness.md |
ATTUNE after a batch to record quality signals, calibrate ranking, and persist reusable patterns. |
reference/claude-code-skills-api.md |
Authoring skill metadata or sandbox rules — frontmatter, routing-sensitive descriptions, dynamic context, install paths. |
reference/claude-md-best-practices.md |
Generating or reconciling CLAUDE.md-adjacent guidance — maturity levels, RFC 2119 wording, split/import decisions. |
reference/cross-tool-rules-landscape.md |
Reconciling project rules across AI tools — CLAUDE.md, .cursorrules, .windsurfrules, AGENTS.md, Copilot instructions. |
reference/meta-prompting-self-improvement.md |
Improving Sigil itself or its long-term calibration loop using self-improvement patterns such as Mistake Ledger and Self-Refine. |
reference/official-skill-guide.md |
Authoring frontmatter, writing descriptions, structuring instructions, or validating against official Anthropic skill standards during CRAFT or VERIFY. |
reference/operating-layer-blueprint.md |
Surveying recurring tasks and selecting the minimum project-local mechanism for blueprint. |
reference/operating-layer-recipe-design.md |
Specifying a repo-tailored single-skill Recipe without moving cross-skill routing out of Nexus. |
reference/operating-layer-workflow-design.md |
Designing pipeline, hub-spoke, hierarchy, or bounded-loop topology for a persistent project workflow. |
reference/operating-layer-handoffs.md |
Handing skill bodies, routing, loops, hooks, and placement to their canonical owners. |
reference/operating-layer-validation.md |
Validating coverage, ownership, topology, mechanism choice, and non-execution boundaries. |
reference/autorun-schema.md |
Emitting the AUTORUN _STEP_COMPLETE block — Sigil-specific Output/Next schema. |
_common/OPUS_5_AUTHORING.md |
Sizing the project skill package or deciding effort allocation across the six-phase pipeline. Critical for Sigil (Knowledge/Meta role): P6, P7. Recommended: P1. |