lubluniky/dale/dale-index

Claude Code and agent skill by lubluniky/dale

Trust tier 3
  • community
Source check pending

Checking that the SKILL.md file is still in the repository.

Create or refresh evidence-backed REQ.md, CONTEXT.md, STATE.md, TDD.md, DESIGN.md, and DECISIONS.md repository primitives. Use for $dale-index, repository mapping, context recovery, or stale project documentation.

Install

npx vskill@latest install lubluniky/dale/dale-index

SKILL.md

Dale Index

Build a compact source of truth from the repository as it exists now. The calling agent integrates and writes; independent discovery uses read-only subagents where repository scope justifies parallel work.

Bound the refresh

Inspect applicable instructions, Git state, manifests, entrypoints, tests, schemas, and existing primitives. Preserve unrelated edits and hand-written content unless current evidence disproves it. No saved Codex project is needed.

For an initial index, cover the repository's material boundaries. For a refresh, use the last documented revision and actual changes to inspect affected claims and their consumers. Include uncommitted changes. When the baseline is missing or unreliable, broaden inspection rather than claiming incremental coverage.

Own each fact once

FileOwns
REQ.mdProduct intent, users, scope, requirements, acceptance criteria
CONTEXT.mdArchitecture, repository map, contracts, constraints, operations
STATE.mdCurrent work, blockers, risks, next actions
TDD.mdTest strategy, actual commands, quality gates, missing coverage
DESIGN.mdExisting visual language, components, states, accessibility
DECISIONS.mdDated decisions, alternatives, evidence, consequences, status

Use the established primitive directory even when its set is incomplete. Add missing documents there and preserve relative links. Use the repository root only when no established location exists; resolve competing locations from repository conventions before writing. Use assets/primitives/ for new files, removing scaffold placeholders. Link to the fact's owning file. Mark non-UI projects explicitly in DESIGN.md. Keep STATE volatile and DECISIONS durable; do not invent dates, owners, product intent, or historical decisions.

Discover and integrate

Work directly for small repositories or a narrow refresh. Delegate independent areas when that saves time or improves coverage; there is no required worker count. Read index-roles.md when splitting discovery. Its roles are coverage suggestions, not mandatory agents.

Use spawn_agent with the minimum context, read-only scope, one evidence output, and no further delegation. Inherit model and effort. Continue a distinct inspection or integration preparation while workers run. Read model-routing.md only for requested routing changes. Separate tasks are optional only when explicitly requested by the user.

Label material claims verified, inferred, unknown, or stale. Source, tests, and runtime observations outrank unsupported documentation; distinguish implemented behavior from exercised behavior. Include compact path/symbol/command evidence.

For consequential contradictions or a broad initial synthesis, use a fresh read-only verifier on raw claims and source evidence. For a small refresh, the coordinator can check the affected claims directly. If independent verification is unavailable, disclose the limitation and finish supported documentation; do not invent a verifier or block accessible indexing merely for tool absence.

Only the coordinator writes primitives. Resolve conflicting claims by evidence quality and freshness, preserve unknowns, and inspect the final diff.

Validate and report

Check that the six documents coexist, referenced paths and commands exist or are labeled unverified, no scaffold remains, and cross-document claims agree. Map known acceptance criteria to available checks. Record revision and coverage so the next refresh can be scoped. Do not run broad builds solely for indexing.

Report created or refreshed documents, material unknowns, and checks performed.

Cancellation and changed scope

On cancellation or revoked scope, interrupt affected workers immediately before other work, invalidate pending outputs, and preserve existing changes. Update or disable related scheduled work only within the user's cancellation scope. For a correction, stop conflicting work first, then send the revised contract; accept subsequent results only against the current request and artifact state.

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

Repository stars
52
Repository forks
3
7-day trend score
20

Stars and forks are for the whole GitHub repository, not this skill alone. Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/lubluniky/dale/dale-index/badge)](https://verified-skill.com/skills/lubluniky/dale/dale-index)

Something wrong with this skill or its page? Tell us.