pensieve

Project knowledge base and workflow router. knowledge/ caches explored file locations, module boundaries, and call chains for reuse; decisions/maxims are established architecture decisions and coding standards to follow, not relitigate; pipelines are reusable workflows; short-term/ stages new conclusions until promotion or deletion. Use self-improve after completing work to capture new insights. Provides seven tools: init, upgrade, migrate, doctor, self-improve, refine, and sync-instructions.

Quality

62%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./SKILL.md

Quality

Content

65%

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

The SKILL.md body is a well-organized, concise router that delegates cleanly to per-tool spec files, but those referenced spec files are absent from the bundle and the routing step itself lacks a request-to-tool matching procedure and validation checkpoints. Actionability and workflow clarity are therefore capped at 2 despite excellent conciseness and good structure.

Suggestions

Include the referenced `.src/tools/*.md` spec files in the bundle so the one-level-deep navigation resolves and actionability reaches copy-paste readiness.

Add a brief decision rule or trigger mapping (which user request maps to which tool) so routing is unambiguous rather than relying on inference.

For destructive/batch tools (migrate, refine, upgrade), note an explicit validate/verify checkpoint — e.g., run `doctor` (read-only) before and after — to lift workflow clarity.

Dimension	Reasoning	Score
Conciseness	The body is lean and efficient — one line per tool with a concrete spec-file pointer, a one-line guardrail ('When in doubt, confirm first.'), and no padding or explanation of concepts Claude already knows, matching the 'lean and efficient; every token earns its place' anchor.	3 / 3
Actionability	Each entry pairs a named action with a concrete file path ('Tool spec: `.src/tools/init.md`'), but the executable detail lives one hop away in `.src/tools/*.md` files that are not present in the bundle, so the body itself is not fully executable/copy-paste ready — matching the 'some concrete guidance but incomplete; missing key details' anchor.	2 / 3
Workflow Clarity	The router logic (match request → read spec → execute) and the 'confirm first' guardrail are clear, but there is no decision procedure for matching requests to tools and no validation/verification checkpoints for destructive/batch tools like migrate, refine, and upgrade, which the guideline says caps workflow clarity at 2.	2 / 3
Progressive Disclosure	Structure and signaling are good — a concise overview with one-level-deep 'Tool spec:' references per tool — but the referenced `.src/tools/*.md` files do not exist in the bundle, so navigation is broken and 'easy navigation' is not achieved, keeping it at 2 rather than 3.	2 / 3
	Total	9 / 12 Passed

Description

60%

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

The description is specific and concrete about what the skill does, but its trigger guidance is narrow (covering only the self-improve sub-tool) and its vocabulary leans toward internal jargon rather than terms users would naturally say. This weakens both completeness and distinctiveness.

Suggestions

Add an explicit overall 'Use when...' clause naming user-facing situations (e.g., 'Use when the user asks to record project knowledge, revisit past decisions, or run a Pensieve maintenance workflow'), not just the self-improve case.

Replace internal jargon ('maxims', 'promotion or deletion', 'call chains') with natural trigger terms a user would actually say, broadening keyword coverage.

Sharpen the niche framing so it is distinguishable from generic memory/knowledge skills — name what is unique to Pensieve up front.

Dimension	Reasoning	Score
Specificity	Names many concrete components and actions — 'caches explored file locations, module boundaries, and call chains', 'established architecture decisions and coding standards', and 'seven tools: init, upgrade, migrate, doctor, self-improve, refine, and sync-instructions' — matching the 'lists multiple specific concrete actions' anchor.	3 / 3
Completeness	The 'what' is clearly stated, but the only explicit 'when' clause ('Use self-improve after completing work to capture new insights') scopes to a single sub-tool rather than the overall skill, leaving the broad trigger only implied; per the guideline a missing/full 'Use when...' clause caps completeness at 2.	2 / 3
Trigger Term Quality	A few natural terms appear ('knowledge base', 'Use self-improve after completing work') but most language is internal jargon ('module boundaries and call chains', 'decisions/maxims', 'promotion or deletion') that users would not naturally say to invoke the skill, matching the 'some relevant keywords but missing common variations' anchor rather than full coverage.	2 / 3
Distinctiveness Conflict Risk	The seven named tools and 'workflow router' framing give it a recognizable niche, but 'project knowledge base' is fairly generic and could overlap with other memory/coding-standard skills, and the weak trigger terms keep it from being clearly conflict-free.	2 / 3
	Total	9 / 12 Passed

Validation

100%

Warnings & errors only

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

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository: kingkongshot/Pensieve
Path: SKILL.md
Commit: 8731f61

Reviewed: 5 days ago

Table of Contents

Discovery Implementation Validation

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.