CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

memory-discipline

Use when deciding whether something learned this session is worth persisting to agent memory, when writing or updating a memory file, and when a saved memory turns out to be wrong. Defines the four memory genres, the selection bar, the evidence-anchor rule, and correction discipline. The repo carries specs/ADRs/beads; memory carries what none of those can.

Quality

87%

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

SKILL.md
Quality
Evals
Security

Quality

Content

82%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.

An unusually tight, opinionated instruction skill: every section states a rule, its failure mode, and a concrete application, with no filler and no over-explanation. The main defect is that its one executable artifact, the referenced memory-lint script, is missing from the bundle, leaving both the lint workflow and the sole external reference dangling.

Suggestions

Ship the referenced 'scripts/memory-lint.bb' in the bundle (or remove/replace the reference with an inline checklist), so the 'Verify links' step is actually executable.

Consider consolidating the write/update path into one short sequenced checklist (decide against the bar → match the genre shape → attach evidence anchors → update the index → lint), pulling together steps currently spread across four sections.

DimensionReasoningScore

Conciseness

Lean and dense throughout: 'Memory is not a diary', 'Every claim carries its anchor, or it's vibes', 'a stale "pending" is a lie with a timestamp'. The few anecdotes (the 30-file corpus, the 28KB file) are evidence anchors for the rules they justify, not padding, and the text never explains concepts Claude already knows.

5 / 5

Actionability

Highly concrete guidance for an instruction-only skill: genre shapes ('the rule → Why: → How to apply: → the incident'), a concrete correction marker ('**CORRECTION (was wrong here):**'), and a size threshold ('past ~10KB ... split'). Not 5 because the one executable command, 'scripts/memory-lint.bb <memory-dir>', references a script that does not exist in this bundle, so it is not copy-paste runnable.

4 / 5

Workflow Clarity

The core decision bar is unambiguous ('its absence already burned a red→green cycle, or would burn one next session') and verification steps are embedded (verify every [[link]] resolves, run the lint script, split files past ~10KB). Not 5 because the writing/updating process is distributed across sections rather than presented as one sequenced workflow with explicit checkpoints.

4 / 5

Progressive Disclosure

Well-organized single file with clear section headers, each appropriately scoped; ~90 lines of cohesive rule content that does not warrant splitting. Minor gap: the sole external reference, 'scripts/memory-lint.bb', does not resolve to a real file in the bundle (no scripts/ directory exists).

4 / 5

Total

17

/

20

Passed

Description

87%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 description with an explicit 'Use when' clause covering three concrete trigger scenarios, a clear statement of what the skill defines, and explicit boundary-drawing against repo-carried records. Only minor gaps: a few natural synonym triggers are missing and the 'what' enumerates framework components rather than actions.

DimensionReasoningScore

Specificity

Lists several specific triggers ('deciding whether something learned this session is worth persisting', 'when writing or updating a memory file', 'when a saved memory turns out to be wrong') and names the concrete artifacts ('the four memory genres, the selection bar, the evidence-anchor rule, and correction discipline'). Not 5 because it enumerates framework components rather than the full set of concrete actions performed, leaving minor coverage gaps.

4 / 5

Completeness

Explicitly answers both: what ('Defines the four memory genres, the selection bar, the evidence-anchor rule, and correction discipline') and when (three concrete 'Use when' trigger clauses). A 4 would require the 'when' to be less explicit, but the trigger phrases here are concrete and specific.

5 / 5

Trigger Term Quality

Natural phrases like 'persisting to agent memory', 'memory file', and 'a saved memory turns out to be wrong' mirror what a user would say. Not 5 because common variations such as 'remember', 'recall', 'MEMORY.md', or 'forget' are absent.

4 / 5

Distinctiveness Conflict Risk

Clear niche with distinct triggers, and it actively disambiguates from adjacent storage ('The repo carries specs/ADRs/beads; memory carries what none of those can'). Minimal conflict risk with other skills.

5 / 5

Total

18

/

20

Passed

Validation

93%

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

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

referenced_paths_exist

Referenced path issues: 1 possibly missing (scripts/memory-lint.bb)

Warning

Total

15

/

16

Passed

Repository

igouss/gitweb-rs

Path

.claude/skills/memory-discipline/SKILL.md

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.