agentsclimarketplace

Lucid

Skill dd3ok/lucid/skills/lucid

Skill-first context hygiene toolkit for AI agents

Install
npx -y skills add dd3ok/lucid --skill lucid

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

One thing to look at

  • 0 stars0 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its author says it does

Copied from the file, not written here

Audits stale, unsafe, contradictory, or bloated agent-facing context and produces cleanup plans. Use for prompt debt, obsolete agent instructions, memory cleanup, source-of-truth drift, AGENTS.md-style docs, and skill descriptions, 프롬프트 부채, and 컨텍스트 정리. Do not use for ordinary README edits, general code refactors, linting, summarization, or creating a memory bank.

SKILL.md

2.3 KB, 460 tokens by cl100k_base, as published. Nobody here has run it

Lucid

Lucid audits agent-facing context debt and creates a cleanup plan.

Use the bundled script for objective checks. Do not rely on chat memory alone to decide what is safe to remove.

Workflow

  1. Resolve this skill directory.

  2. Run scan from the target repository root:

    python3 <lucid-skill-dir>/scripts/lucid.py scan --root . --format json

  3. Run audit:

    python3 <lucid-skill-dir>/scripts/lucid.py audit --root . --format json --out .lucid/audit.json

  4. Create a cleanup plan:

    python3 <lucid-skill-dir>/scripts/lucid.py plan --root . --audit .lucid/audit.json --out .lucid/plan.md

  5. Optionally generate a review-only patch suggestion:

    python3 <lucid-skill-dir>/scripts/lucid.py suggest --root . --audit .lucid/audit.json --out .lucid/suggested.patch

  6. Classify every item as one of:

    remove
replace-with-pointer
move-to-reference
move-to-validator
move-to-eval
keep-with-reason
manual-review

  7. Do not edit or delete files unless the user explicitly asks to apply the plan.

  8. After approved edits, run:

    python3 <lucid-skill-dir>/scripts/lucid.py verify --root . --strict

Safety

Lucid is read-only by default.

Do not delete tracked files wholesale. Do not remove schema, migration, alias, marker, protocol, or compatibility fields just because they look old.

Do not preserve obsolete names in user-facing context merely to forbid them. Prefer source-of-truth pointers, validators, or evals over natural-language warnings.

Before memory cleanup, old-looking compatibility-field removal, or negative-warning cleanup, read references/memory-retention-rubric.md, references/compatibility-safety.md, or references/negative-residue.md, respectively.

Keep looking

Skills are one crate of 328,083. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.