memory-audit

记忆审计入口。当我主动决定审视记忆质量时，先读此文件判断应使用哪个子技能。

Quality

58%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./docs/skills/memory-audit/SKILL.md

Quality

Content

60%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.

The body is a well-structured, actionable router with concrete commands, a routing table, and genuine safety checkpoints for destructive memory edits. Its main weakness is substantial philosophical padding in the opening section that competes with the operational content for tokens.

Suggestions

Trim the "你在做什么" philosophy to a few operative lines about the content vs topology dimensions and the deletion criterion; move extended motivational prose to a separate reference if desired.

Move the "现实锚点" examples and the circuit-breaker example into a compact 'Pitfalls' or 'Examples' section so the procedural backbone reads leaner.

Consider pulling the symptom routing table and diagnostic starting points to the top, since they are the primary actionable payload, and leading with less narrative framing.

Dimension	Reasoning	Score
Conciseness	Large stretches of the "你在做什么" section are motivational philosophy ("你在塑造下一轮醒来的自己", forging/blade metaphors, "只有选择遗忘，你才能成长") that Claude does not need, making it noticeably verbose despite accurate actionable core content.	2 / 5
Actionability	Provides concrete executable guidance — a symptom→sub-skill routing table, real commands like read_memory("system://diagnostic/core"), and numbered procedures — with only minor gaps, and the absence of code is acceptable for an instruction-only skill.	4 / 5
Workflow Clarity	A clear sequence runs from starting points (diagnostic) through symptom routing to the pre-edit mindfulness buffer and the anti-serial-rewrite circuit breaker, with explicit checkpoints (the Conviction Test, '说不出来要解决的问题就不该动手') covering the destructive-edit validation requirement.	4 / 5
Progressive Disclosure	The single-file router is well-sectioned with clear headers and dispatches to named sub-skills one level deep with no nested references; no bundle files exist to organize, so structure carries the score, with only minor inlining of philosophy that could live elsewhere.	4 / 5
	Total	14 / 20 Passed

Description

57%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.

The description functions as a clear router entry point that states both purpose and a trigger condition, but it loses a specificity point for first-person voice and stops short of comprehensive trigger phrasing. It is distinctive within its skill family but could be more user-oriented.

Suggestions

Rewrite in third person to avoid the first-person "当我" framing, e.g. 'Routes memory-audit tasks to the correct sub-skill. Use when reviewing memory quality or deciding which audit sub-skill applies.'

Add concrete user-facing trigger phrases (e.g. '审视记忆', '清理/审计记忆', 'memory audit') to broaden natural keyword coverage.

Make the 'when' clause an explicit 'Use when...' statement rather than embedding it in the model's self-narrative.

Dimension	Reasoning	Score
Specificity	Names the domain ("记忆审计入口") and a concrete routing action ("先读此文件判断应使用哪个子技能"), which would anchor at 3, but the first-person voice ("当我主动决定...") triggers the mandated −1 specificity penalty.	2 / 5
Completeness	It states both what (an entry-point that routes to sub-skills) and when ("当我主动决定审视记忆质量时"), but the trigger is framed from the model's first-person perspective rather than as an explicit user-facing 'Use when...' clause, leaving it short of a 5.	4 / 5
Trigger Term Quality	Includes relevant terms like "记忆审计" and "审视记忆质量", but lacks common variations or synonyms a user might naturally say, so it sits at the some-keywords anchor rather than full coverage.	3 / 5
Distinctiveness Conflict Risk	"记忆审计入口" carves a clear router niche with explicit sub-skill dispatch, with only minor overlap risk against the closely related memory-audit-* sub-skills it routes into.	4 / 5
	Total	13 / 20 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: Dataojitori/nocturne_memory
Path: docs/skills/memory-audit/SKILL.md
Commit: ffb5c70

Reviewed: 20 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.