CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

convex-improve-convex-plugin

Send this coding session's transcript to the Convex team for an AI post-mortem that improves the quickstart system. Run only when the user explicitly asks.

Quality

83%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide
SecuritybySnyk

Critical

Do not install without reviewing

SKILL.md
Quality
Evals
Security

Quality

Content

92%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 an exemplary lean, actionable workflow: one command, explicit consent gating with a retry loop, output markers, and error handling, all in under 50 well-organized lines. The only notable gap is the unresolved '<anteater>' base-URL placeholder, which keeps the command from being fully copy-paste executable. No bundle files exist, so the structure is appropriately self-contained.

DimensionReasoningScore

Conciseness

The body is lean and instruction-dense: a one-line purpose, a four-step workflow with the exact command and output markers, and four terse rules, with no explanation of concepts Claude already knows. Every line carries operational information, matching 'Lean and efficient; assumes Claude's competence; every token earns its place'.

5 / 5

Actionability

The guidance is mostly executable: a concrete curl command with argument syntax, the consent re-run invocation ('re-run appending --consent always|once|never'), specific exit code and markers to watch, and a defined summary format. It falls short of fully copy-paste ready because the '<anteater>' URL placeholder and the undefined base endpoint leave a gap Claude cannot resolve from the skill alone, matching 'concrete code or commands with minor gaps'.

4 / 5

Workflow Clarity

The four steps are clearly sequenced with explicit checkpoints and a feedback loop: CONSENT_REQUIRED with exit 4 halts sending until the user answers and triggers a re-run with --consent, and REVIEW_SOURCE/REVIEW_SUBMITTED/REVIEW_DONE markers track progress with a defined terminal state. Error recovery for REVIEW_NO_TRANSCRIPT is also specified, matching 'Clear sequence with explicit validation steps; feedback loops for error recovery'.

5 / 5

Progressive Disclosure

This is a short single-purpose skill (under 50 lines) with no bundle files (no references/, scripts/, or assets/ exist), and its content fits entirely in well-organized sections (purpose, Workflow, Rules) with no material that belongs in separate files. Per the rubric's simple-skill guidance, this earns 5 with just well-organized sections.

5 / 5

Total

19

/

20

Passed

Description

66%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 clear, honest, and highly distinctive with an explicit trigger gate, but it is thin on capability detail and natural trigger vocabulary. Its 'when' guidance is a gate rather than concrete phrases users would naturally say. A modest rewrite adding what findings the post-mortem produces and how a user would phrase the request would lift it to top-tier quality.

Suggestions

Expand specificity by naming the concrete capabilities of the post-mortem, e.g., 'Returns structured findings (ambiguous instructions, agent-stuck patterns, tooling failures) for the runbook, bootstrap script, and skills' — this converts a single pipeline action into several specific ones.

Add natural trigger phrases to the when-clause, e.g., 'Run only when the user explicitly asks to share the session transcript, run a post-mortem, or send feedback to the Convex team' — this supplies the vocabulary users would actually say.

Include common synonyms such as 'session review', 'share transcript', or 'Convex feedback' so the skill matches varied user phrasings rather than only exact requests.

DimensionReasoningScore

Specificity

The description names a clear domain and 1-2 concrete actions — "Send this coding session's transcript to the Convex team" for "an AI post-mortem that improves the quickstart system" — but does not enumerate several specific actions, matching the 'Names domain and 1-2 concrete actions' anchor. It is below score 4 because a single pipeline action (send transcript for post-mortem) is described rather than a list of specific capabilities, and above score 2 because the actions are concrete rather than generic.

3 / 5

Completeness

Both parts are explicitly present: the 'what' ('Send this coding session's transcript to the Convex team for an AI post-mortem that improves the quickstart system') and the 'when' ('Run only when the user explicitly asks'), which qualifies as equivalent explicit trigger guidance so the missing-'Use when' cap does not apply. It falls short of score 5 because the when-clause lacks concrete trigger phrases, exactly matching 'Has both what and when; when could be more explicit or specific'.

4 / 5

Trigger Term Quality

Relevant domain keywords like "transcript", "post-mortem", and "Convex team" are present and would be used by a user requesting this, but common variations (e.g., "share session", "review my session", "feedback") are absent, matching 'Some relevant keywords but missing common variations'. It is not score 4 because the only when-clause ('Run only when the user explicitly asks') is a gating statement that contributes no natural trigger phrases a user would say.

3 / 5

Distinctiveness Conflict Risk

The niche is unmistakable — sending the session transcript to the Convex team for a post-mortem of the quickstart system — and the 'Run only when the user explicitly asks' gate makes accidental triggering for another skill extremely unlikely, matching 'Clear niche with distinct triggers; minimal conflict risk'. Score 4 would require acknowledged overlap with closely related skills, which the highly specific scope and explicit-request gate preclude.

5 / 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

openclaw/clawhub

Path

.agents/skills/convex-improve-convex-plugin/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.