CtrlK
BlogDocsLog inGet started
Tessl Logo

self-evolving-single-agent

Use when generating a single installable agent that should keep learning, track sources, refresh research, propose repairs, or improve itself over time without becoming a multi-agent team.

71

Quality

86%

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 self-evolving-single-agent in agentlas-ai/Agentlas-OS

SKILL.md
Quality
Evals
Security

Quality

Content

86%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 content is a concise, well-structured instruction skill with concrete deliverables and a clear sequence including an approval gate. The main gap is the absence of explicit validate/retry feedback loops and fully copy-paste-ready commands, which keep actionability and workflow clarity just below the top anchor.

Suggestions

Add an explicit validation/verification checkpoint in the procedure (e.g. 'After drafting the worker prompt, run the capability-eval-plan and only proceed when it passes') to strengthen feedback loops.

Include one concrete example command or snippet for invoking the research gate or emitting a self-evolution proposal, so the guidance is copy-paste ready rather than referencing external contracts.

Make the approval-gate loop explicit with a 'draft -> await approval -> apply -> re-verify' sub-sequence in step 6.

DimensionReasoningScore

Conciseness

The body is a lean numbered procedure plus an Output list with no concept padding or over-explanation; every step is actionable instruction that assumes Claude's competence, matching the top anchor.

5 / 5

Actionability

Concrete guidance names exact artifacts ('.agentlas/memory-map.json', 'docs/builder-interview.md') and specifics ('8-12 question first batch'), but execution leans on referenced contracts and produces no copy-paste commands, leaving minor gaps versus the fully-executable top anchor.

4 / 5

Workflow Clarity

Seven clearly sequenced steps with an explicit approval-gate checkpoint (step 6) and a required 'verification' output; it lacks explicit validate-and-retry feedback loops within the procedure, so it is just below the top anchor.

4 / 5

Progressive Disclosure

Under 50 lines with no bundle files needed; well-organized into Procedure and Output sections, qualifying for the top anchor under the simple-skill exception. The referenced .agentlas/docs paths are artifacts to create, not navigation references.

5 / 5

Total

18

/

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.

The description is explicit, third-person/imperative in voice, and answers both 'what' and 'when' with a distinctive niche. It is strong overall, with only slightly abstract capability verbs keeping specificity and trigger coverage just short of the top anchor.

DimensionReasoningScore

Specificity

Lists several specific capabilities ('keep learning, track sources, refresh research, propose repairs, or improve itself over time'), but these are behavior-level verbs rather than the concrete operations (e.g. 'extract text', 'fill forms') the top anchor requires, so it sits just below 5.

4 / 5

Completeness

Explicit 'Use when generating a single installable agent that should keep learning...' states both what the skill produces and when to invoke it with concrete trigger phrasing, matching the top anchor.

5 / 5

Trigger Term Quality

Natural phrases a user would say appear ('generating a single installable agent', 'improve itself', 'propose repairs', 'refresh research'); coverage is good but a few common synonyms or variants are missing, placing it above the midpoint but not at comprehensive.

4 / 5

Distinctiveness Conflict Risk

The 'single installable agent ... without becoming a multi-agent team' framing carves a clear niche distinct from multi-agent team skills, with minimal overlap risk.

5 / 5

Total

18

/

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
agentlas-ai/Agentlas-OS
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.