CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

convex-backup

Set up Convex backups and run a restore DRILL that proves recovery — snapshot, restore into a throwaway preview, assert the data came back — plus a schedule matched to your RPO and a gated recovery runbook.

Quality

82%

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 convex-backup in openclaw/clawhub

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.

An exemplary operational skill: concrete commands for every step, explicit restore verification with failure semantics, safety guardrails (throwaway preview, hygiene, deploy-guard gating), and clean structure with no padding beyond minor Workflow/Rules duplication. The only refinements would be an explicit remediation path when the assert fails and consolidating the duplicated Rules bullets.

DimensionReasoningScore

Conciseness

The body is dense and operational with no explanations of concepts Claude already knows — every line carries instructions, commands, or safety constraints. It sits at anchor 4 rather than 5 because the Rules section duplicates several Workflow points (0-rows-is-failure, RPO schedule, hygiene, preview-not-prod), and asides like "(the half almost nobody does — this is the point)" could be trimmed.

4 / 5

Actionability

Every step has copy-paste-ready commands: "npx convex export --path backup-<date>.zip", "npx convex deploy --preview-create restore-drill-<date>", "npx convex import backup-<date>.zip --deployment restore-drill-<date> --replace", and the prod runbook "npx convex import backup.zip --replace --prod". Verification names specific MCP tools (tables, data, runOneoffQuery), preconditions and a fallback path are given, and even a CLI gotcha ("there is no --preview-name on import") is flagged — this matches the fully-executable anchor 5.

5 / 5

Workflow Clarity

The six-step sequence is explicit with a leading GUARD, a precondition check with a stated fallback, and a genuine validation checkpoint (ASSERT recovery: row counts + real-record sample, with a 0-rows restore defined as FAILED). It falls short of anchor 5 only because there is no error-recovery feedback loop — the workflow says to surface a failed drill loudly but not what to diagnose or retry when the assert fails.

4 / 5

Progressive Disclosure

The skill is under 50 lines, needs no external references (none exist in the bundle), and is cleanly organized into an intro, a numbered Workflow, and Rules — squarely matching the guideline that short single-purpose skills with well-organized sections and no needed external files score 5 on progressive disclosure.

5 / 5

Total

18

/

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 specific, distinctive description with strong concrete verbs and good natural keywords, weakened mainly by the absence of any "Use when…" trigger guidance and by imperative/second-person phrasing instead of third person. Adding an explicit trigger clause and a couple of synonyms (export, disaster recovery) would lift it substantially.

Suggestions

Append an explicit trigger clause, e.g. "Use when the user asks to back up, export, restore, or recover a Convex deployment, or wants disaster-recovery confidence for their Convex data."

Add the natural synonym "export" (the actual CLI verb) and "disaster recovery" to the trigger terms so users phrasing the request that way still match.

Rewrite in third person ("Sets up Convex backups and runs a restore drill…") to match the expected description voice.

DimensionReasoningScore

Specificity

The description lists multiple concrete actions — "Set up Convex backups and run a restore DRILL that proves recovery — snapshot, restore into a throwaway preview, assert the data came back — plus a schedule matched to your RPO and a gated recovery runbook" — which matches the anchor 5 example of comprehensive specific actions. However, it is written in imperative/second-person voice ("Set up…", "your RPO") rather than third person, which the guidelines penalize by reducing specificity by 1.

4 / 5

Completeness

The "what" is clear and detailed (backup setup, restore drill, schedule, runbook), but there is no "Use when…" clause or any equivalent explicit trigger guidance. Per the judging guidelines, a missing trigger clause caps completeness at 3, which fits the anchor: clear what, when absent.

3 / 5

Trigger Term Quality

Solid natural keywords: "backups", "restore", "recovery", "snapshot", "RPO", "runbook" — phrases a user would plausibly say when needing this skill. A few natural variants are missing, notably "export" (the actual command users associate with Convex backups) and "disaster recovery", so it sits between anchors 3 and 5 rather than clearly at 5.

4 / 5

Distinctiveness Conflict Risk

"Convex backups" plus restore-drill/recovery framing carves out a clear niche with distinct triggers and minimal conflict risk — it would not fire for generic backup requests or for Convex schema-migration work (the body's migrate-rehearse is not mentioned in the description). This clearly matches the anchor 5 example of a distinct, low-conflict description.

5 / 5

Total

16

/

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

FacundoLucci/autonomous-buyer

Path

.claude/skills/convex-backup/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.