CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

convex-self-heal

Production error → triaged, root-caused, repaired, and certified (tsc + rehearsal + reproduce-then-gone) fix PR for a human to merge — then confirm the error stops recurring. Never auto-merges.

Quality

76%

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 ./.agents/skills/convex-self-heal/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

77%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 lays out a rigorous, well-gated production workflow with explicit certification checkpoints, stop conditions, and a reopen loop — model sequencing for a risky operation. Its weaknesses are token redundancy (the Rules section restates the workflow, and the intro is competitive positioning rather than instruction) and a couple of unspecified details like the append-only audit record.

Suggestions

Cut or compress the Sentry/Datadog/Vercel positioning paragraph to a single sentence; it does not change execution behavior.

Trim the Rules section to only constraints not already stated in the Workflow steps (e.g., keep never-auto-merge and stop-if-unclear, drop the re-statements of certify/triage/branch/promote).

Specify the append-only audit record concretely — where it lives and what each entry contains — or point to where that detail is defined.

DimensionReasoningScore

Conciseness

The body is mostly efficient and assumes domain competence, but the opening positioning paragraph ('Sentry/Datadog/Vercel can go error→investigate→draft-PR, but they treat the backend as opaque...') and a Rules section that largely restates workflow steps 1–9 are unnecessary or duplicated. Not 2: there is no padding explaining concepts Claude already knows; not 4: the duplication and marketing-style intro exceed minor trimmable instances.

3 / 5

Actionability

Most steps name the exact capability and command to invoke — 'tsc --noEmit clean', 'data/run-once-query over the sentinel table, or the monitor's prod_error event', 'migrate-rehearse on a preview seeded with a prod snapshot' — making the orchestration executable. Not 5: a few key details are unspecified, such as where the append-only audit record lives and what a PR must contain structurally; not 3: concrete commands and tool selections cover nearly every step.

4 / 5

Workflow Clarity

The 9-step sequence has explicit validation checkpoints — the certification checklist (a)–(d), stop conditions ('If root cause is unclear, STOP and report', 'A fix that fails any applicable certification is NOT proposed'), and a reopen feedback loop if the error recurs. Not 4: validation and error-recovery loops are fully explicit for this high-stakes, prod-touching workflow.

5 / 5

Progressive Disclosure

The body is a 38-line single file with well-organized 'Workflow' and 'Rules' sections and no bundle files, none of which are needed at this size. Per the under-50-line guideline, well-organized sections alone warrant a 5.

5 / 5

Total

17

/

20

Passed

Description

63%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 precisely enumerates the capability's pipeline with concrete certification steps and an explicit safety boundary, but it entirely lacks a 'Use when' trigger clause and favors compressed internal jargon over the natural keywords a user would actually say. It scores well on what it does and poorly on when to reach for it.

Suggestions

Add an explicit trigger clause, e.g. 'Use when a production error appears or recurs, or when the user asks for an automated/self-healing fix loop or a fix PR for a prod incident.'

Include natural user phrasings and synonyms as trigger terms — 'production bug', 'incident', 'Sentry/Datadog alert', 'recurring error', 'debug and fix prod' — instead of internal jargon like 'reproduce-then-gone' and 'rehearsal'.

Keep the safety boundary ('Never auto-merges') but place it after the trigger clause so the when-to-use question is answered first.

DimensionReasoningScore

Specificity

The description enumerates the entire pipeline concretely — 'triaged, root-caused, repaired, and certified (tsc + rehearsal + reproduce-then-gone) fix PR for a human to merge — then confirm the error stops recurring' — covering every stage from error intake to post-deploy confirmation. Not 4: there are no coverage gaps; each action is explicit rather than implied.

5 / 5

Completeness

The 'what' is fully explicit (end-to-end fix pipeline with certification and a human merge gate), but there is no 'Use when...' clause or any equivalent explicit trigger guidance, which caps completeness at 3 per the judging guidelines. Not 2: the 'what' is clear and specific, not vague.

3 / 5

Trigger Term Quality

It includes some natural phrases users would say ('production error', 'fix PR', 'stops recurring'), but leans on internal jargon ('triaged', 'certified', 'reproduce-then-gone', 'rehearsal') and misses common variations like 'bug', 'incident', 'debug', 'recurring error', or monitoring-tool names. Not 4: key synonyms and natural phrasings are absent.

3 / 5

Distinctiveness Conflict Risk

The certification-plus-human-merge framing ('certified... fix PR for a human to merge — Never auto-merges') carves a distinct niche unlikely to fire for unrelated skills. Not 5: generic 'production error → fix' phrasing could still overlap with general debugging or incident-response skills.

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

openclaw/clawhub

Path

.agents/skills/convex-self-heal/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.