CtrlK
BlogDocsLog inGet started
Tessl Logo

ai-docs

Generate, audit, and compress AI-optimized documentation for codebases. This skill applies research-backed principles from three peer-reviewed papers to create documentation that maximizes AI agent performance instead of degrading it. Use this skill whenever the user asks to "write docs", "create a CLAUDE.md", "document a module", "audit documentation", "compress docs", "optimize docs for AI", "write module docs", "create domain documentation", "review doc quality", or mentions documentation bloat, token waste, or AI context efficiency. Also trigger when the user references "ai-docs principles", "telegram style docs", or wants to apply the "Lost in the Middle", "Less is More", or "AGENTS.md" research findings.

72

Quality

88%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Passed

No findings from the security scan

The canonical home for this skill is ai-docs in alpacapurpura/luana-method

SKILL.md
Quality
Evals
Security

Quality

Content

77%Weight 40%Scale 1-5

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

A well-structured, lean skill body that turns research principles into four concrete operating modes with a final validation checklist. It appropriately offloads detail to two real reference files and largely assumes Claude's intelligence; the main gaps are missing explicit feedback loops for the destructive compress/delete steps.

Suggestions

Add an explicit validate->fix->retry feedback loop to Mode 3 (Compress): after deleting lines, re-check the survivor doc still satisfies the Quality Checklist before reporting the reduction percentage.

Trim explanations of concepts Claude already knows — e.g., the one-line rationale after each Anti-Patterns row and 'LLMs have U-shaped attention' — since the body's own Law 3 already conveys the rule.

Consider moving the full audit-report code block into references/ to keep SKILL.md as a pure overview, or condense it to a compact schema so the inlined template does not border on detail that belongs in a reference file.

DimensionReasoningScore

Conciseness

The body is efficient and assumes Claude's competence (no library tutorials, no explanation of what a PDF is), though a few lines restate knowledge Claude has (e.g., 'LLMs have U-shaped attention' and the Anti-Patterns rationales) that could be trimmed; matches 'efficient; minor instances of over-explanation'.

4 / 5

Actionability

Provides concrete, copy-paste-ready artifacts — the audit-report code block with placeholders and specific line-target rules (80/100/150) — but Mode 1 and Mode 3 lean on instruction-style guidance rather than executable commands, leaving minor gaps.

4 / 5

Workflow Clarity

Each mode is a clearly numbered sequence and Mode 2 includes an explicit deliverable template plus the final Quality Checklist acts as a validation gate; it stops short of explicit error-recovery feedback loops (validate -> fix -> retry) for the destructive delete operations in Mode 3.

4 / 5

Progressive Disclosure

Body is an overview that signals one-level-deep references clearly ('read references/research-foundations.md', 'read references/templates.md'), both of which exist as real bundle files; the audit report template is arguably inlinable detail but is well-placed, so minor organization gaps hold it at 4 rather than 5.

4 / 5

Total

16

/

20

Passed

Description

100%Weight 40%Scale 1-5

Based on the skill's description, can an agent find and select it at the right time? Clear, specific descriptions lead to better discovery.

A strong, trigger-rich description that crisply states what the skill does and enumerates a wide range of natural user phrases that should activate it. It avoids vague fluff and third-person voice is maintained throughout.

DimensionReasoningScore

Specificity

Lists multiple concrete actions — 'Generate, audit, and compress AI-optimized documentation' — and each mode (generate/audit/compress/map) maps to specific operations, giving comprehensive coverage of the skill's capabilities.

5 / 5

Completeness

Explicitly answers 'what' (generate/audit/compress AI-optimized docs) and 'when' via an explicit 'Use this skill whenever the user asks to...' clause with concrete trigger phrases, hitting the top anchor.

5 / 5

Trigger Term Quality

Includes a dense set of natural phrasings users would say ('write docs', 'create a CLAUDE.md', 'audit documentation', 'compress docs', 'token waste', 'AI context efficiency') plus research name-drops as synonyms, matching the comprehensive-coverage anchor.

5 / 5

Distinctiveness Conflict Risk

The niche (AI-optimized documentation with named research principles like 'Lost in the Middle'/'AGENTS.md') and its trigger phrases ('telegram style docs', 'ai-docs principles') are distinctive and unlikely to collide with unrelated skills.

5 / 5

Total

20

/

20

Passed

Validation

100%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository
alpacapurpura/ap_sales_agent
Reviewed

Table of Contents

Is this your skill?

If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.