CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

self-improvement

Captures learnings, errors, corrections, and feature requests to enable continuous improvement. Use when: (1) User corrects Claude ('No, that's wrong...', 'Actually...'), (2) User requests a capability that doesn't exist, (3) Claude realizes its knowledge is outdated or incorrect, (4) A better approach is discovered for a recurring task, (5) Receiving a Handoff block from self-healing (a recurring verified heal with Recurrence-Count at least 3) to distill into a memory file or new skill. For ACTIVE runtime failures where the agent needs to apply and verify a fix mid-task, use `self-healing` instead (it files HEAL- entries with proof; self-improvement promotes accumulated patterns). Also review learnings before major tasks. For CI-only/headless learning capture, use self-improvement-ci.

0.89x
Quality

85%

Does it follow best practices?

Impact

84%

0.89x

1 of 3 eval scenarios. Add 2 more for a full score.

SecuritybySnyk

Passed

No findings from the security scan

The canonical home for this skill is self-improvement in pskoett/pskoett-ai-skills

SKILL.md
Quality
Evals
Security

Quality

Content

75%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 strong, highly actionable instruction skill — complete entry templates, executable grep commands, and well-sequenced promotion/extraction workflows with approval gates and verification steps. Its weaknesses are token efficiency (~500 lines with redundant priority/area tables and overlapping best-practices sections) and an orphaned references/examples.md that is never linked from the main file.

Suggestions

Trim the Priority Guidelines and Area Tags tables (their values are already enumerated inline in the entry templates) and cut or merge the Best Practices and Multi-Agent Support sections, which largely restate Detection Triggers.

Link references/examples.md from SKILL.md (e.g., a '## Examples' pointer) so its 383 lines of example entries are discoverable, or fold its essential examples into the entry templates and drop the file.

Add brief failure-path guidance to the Promotion and Extraction workflows (e.g., what to do when the fresh-session verification fails or when eligibility checks reject a candidate), completing the feedback loops for workflow clarity.

DimensionReasoningScore

Conciseness

The body is dense and specific (no concept-explanation fluff), but at ~500 lines it includes unnecessary material: the Priority Guidelines and Area Tags tables restate the enum values already inline in the entry templates ("low | medium | high | critical", "frontend | backend | infra | tests | docs | config"), and the Best Practices and Multi-Agent Support sections restate guidance covered by Detection Triggers. Not 4: the redundancy is more than minor trimmable instances; not 2: most sections are actionable and free of padded explanation.

3 / 5

Actionability

Fully executable throughout: complete copy-paste markdown templates for all three entry types, exact ID format with examples ("LRN-20250115-001"), runnable bash ("grep -h \"Status\**: pending\" .learnings/*.md | wc -l", "./skills/self-improvement/scripts/extract-skill.sh skill-name --dry-run"), and a concrete promotion diff example. Matches the fully-executable anchor covering common cases.

5 / 5

Workflow Clarity

Multi-step workflows are clearly sequenced with checkpoints: promotion step 3 "Show the proposed diff for human approval before writing instructions", extraction step 5 "Verify: Read skill in fresh session to ensure it's self-contained", and a Skill Quality Gates checklist. Not 5: no error-recovery feedback loops (e.g., what to do when the fresh-session verification or a promotion eligibility check fails), and Periodic Review actions are an unordered bullet list; not 3: validation checkpoints are explicit rather than missing.

4 / 5

Progressive Disclosure

Good bundle structure with clearly signaled, purpose-scoped one-level-deep references: "For installation commands, see references/installation.md... Read these only when installing or configuring the skill", plus hooks-setup.md, amp-setup.md, context-maintenance.md, assets/SKILL-TEMPLATE.md, and scripts/extract-skill.sh — all real files. Not 5: references/examples.md (383 lines) is orphaned — never linked from SKILL.md, so undiscoverable — and inline enum tables plus Gitignore Options could live in a reference to slim the overview; not 3: the split between overview and detail is deliberate and mostly well-placed.

4 / 5

Total

16

/

20

Passed

Description

92%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 is highly effective: third-person, explicit what-and-when structure with five enumerated triggers, quoted user phrases, and explicit disambiguation from self-healing and self-improvement-ci. Its only weakness is length and a few missing natural trigger synonyms (e.g., "remember this", "lesson learned"), which keeps trigger term quality at 4 rather than 5.

DimensionReasoningScore

Specificity

"Captures learnings, errors, corrections, and feature requests" plus "distill into a memory file or new skill", "promotes accumulated patterns", and "review learnings before major tasks" name multiple concrete actions covering all of the skill's workflows (capture, promotion, review). Not 4: the action list covers the skill comprehensively with no material gaps; not below 5 because every named action is concrete rather than generic.

5 / 5

Completeness

Explicitly answers what ("Captures learnings, errors, corrections, and feature requests to enable continuous improvement") and when ("Use when: (1)...(5)" with concrete quoted trigger phrases), plus when-not guidance ("For ACTIVE runtime failures... use `self-healing` instead"). Matches the anchor for clear, explicit what AND when with concrete trigger phrases; neither neighbor anchor fits better.

5 / 5

Trigger Term Quality

Includes natural user phrases such as "No, that's wrong...", "Actually...", "User requests a capability that doesn't exist", and "knowledge is outdated or incorrect" — good coverage of natural triggers. Not 5: common variations users would say are missing (e.g., "remember this", "save this learning", "lesson learned", "I keep hitting this"); not 3: the quoted correction phrases and capability-request phrasing go well beyond a bare domain keyword.

4 / 5

Distinctiveness Conflict Risk

Maintains a clear niche and actively disambiguates the two highest-risk siblings: "use `self-healing` instead (it files HEAL- entries with proof; self-improvement promotes accumulated patterns)" and "For CI-only/headless learning capture, use self-improvement-ci". Not 4: the explicit boundary marking against closely related skills leaves minimal conflict risk rather than minor overlap.

5 / 5

Total

19

/

20

Passed

Validation

87%

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

Validation — 14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

skill_md_line_count

SKILL.md is long (508 lines); consider splitting into references/ and linking

Warning

relative_links

Relative link issues: 1 suspicious

Warning

Total

14

/

16

Passed

Repository

pskoett/pskoett-ai-skills

Path

.agents/skills/self-improvement/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.