CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

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.

Quality

71%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Low

Low-risk findings worth noting

Fix and improve this skill with Tessl

tessl review fix ./skills/self-evolving-single-agent/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

57%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 compact and efficiently written with a sensible procedure outline, but it operates at the naming level: referenced artifacts have no templates or examples, verification is asserted as an output rather than specified as a step, and a referenced contract file is absent from the bundle. It reads as a checklist of things to create rather than executable guidance for creating them.

Suggestions

Add a short template or worked example for at least the builder-interview question batch and the prompt-performance contract so steps 2 and 5 are executable rather than just named.

Specify the concrete verification procedure (e.g., how to run .agentlas/capability-eval-plan.json checks, confirm memory files exist, validate the global command) with a fix-and-retry loop after generation.

Either bundle contracts/builder-interview-research-gate.md or inline its procedure in the body — the path is referenced as something to 'run' but does not exist in the skill directory.

DimensionReasoningScore

Conciseness

The ~44-line body is lean with no explanations of concepts Claude already knows, but the 'File revision approval' section is jargon-dense ('PM Soul/project owner', 'revision-v1/evolution-v2 host capabilities', 'legacy chip') in a way that could be tightened.

4 / 5

Actionability

Names specific files to create ('docs/builder-interview.md', '.agentlas/memory-map.json') and one concrete parameter ('8-12 question first batch'), but provides no templates, examples, or criteria for what those artifacts should contain — the guidance names outputs without fully instructing how to execute each step.

3 / 5

Workflow Clarity

A clear numbered 1-7 sequence exists and a proposal-first approval gate is stated, but there is no explicit validation/verification procedure or fix-and-retry loop; since the workflow changes tools, connectors, secrets, and core instructions, the missing-validation cap holds it at 3 despite 'verification' being listed as an output.

3 / 5

Progressive Disclosure

Sections are well-organized for a short skill, but step 2 directs 'Run contracts/builder-interview-research-gate.md' — a path that does not exist in the skill directory (no bundle is provided) — a dangling reference that is neither clearly signaled nor navigable.

3 / 5

Total

13

/

20

Passed

Description

70%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 well-formed description with an explicit trigger clause and a clear single-vs-multi-agent boundary, weakened slightly by feature-level rather than concrete action language and missing common synonyms like 'self-improving'. It clearly beats the vague examples and lands just below the top anchors.

DimensionReasoningScore

Specificity

Names the domain ('generating a single installable agent') and lists capability-level actions ('track sources, refresh research, propose repairs'), but these describe features rather than concrete operations — 'keep learning' and 'improve itself' are abstract, so it does not reach the 'several specific actions, minor gaps' bar of anchor 4.

3 / 5

Completeness

An explicit 'Use when generating...' clause provides a clear 'when', and the 'what' (generates such a self-evolving agent package) is present but merged into the same sentence rather than clearly separated with concrete trigger phrases as anchor 5 requires.

4 / 5

Trigger Term Quality

Good natural-term coverage ('single installable agent', 'keep learning', 'improve itself', 'multi-agent team') that a user asking for a self-improving agent would plausibly say, though common synonyms like 'self-improving' or 'autonomous agent' are missing.

4 / 5

Distinctiveness Conflict Risk

'Without becoming a multi-agent team' carves an explicit boundary from multi-agent skills, leaving only minor overlap risk with general agent-generation skills — mostly distinct rather than a fully clear niche.

4 / 5

Total

15

/

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.

Validation — 16 / 16 Passed

Validation for skill structure

No warnings or errors.

Repository

agentlas-ai/Agentlas-OS

Path

skills/self-evolving-single-agent/SKILL.md

Revision

2b4e27b

Reviewed

Table of Contents

DiscoveryImplementationValidation

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.

Claim skill