CtrlK
BlogDocsLog inGet started
Tessl Logo

meta-long-running-build-watchdog

[DEPRECATED] Build watchdog — launches arbitrary commands from the user message in tmux and lets sub-agent auto-apply a fix. Disabled pending the E5 bounded sub-agent contract + Jinja sandbox + side-effect ledger (plan §3.1 A1/A8 / §5.3 E4): the launch task interpolates raw user_message into a shell-bound tmux session and the heal step lets sub-agent mutate state with no rollback. Do not re-enable without `metadata.opensquilla.risk: high` + capabilities {shell, tmux, filesystem-write, subprocess} and a saga-style compensation step.

49

Quality

55%

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

Fix and improve this skill with Tessl

tessl review fix ./src/opensquilla/skills/exp/meta-long-running-build-watchdog/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

58%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 commendably concise and appropriately structured for a simple skill, but it provides only vague, non-executable guidance and lacks the validation/feedback checkpoints a state-mutating heal workflow requires.

Suggestions

Add concrete, copy-paste-ready commands for the fallback (e.g. the actual `tmux new-session -d ...`, capture-pane, and memory_save invocations) instead of prose hints.

Include an explicit validation/feedback checkpoint before the heal step records to memory (e.g. verify the proposed fix compiles or the build re-runs green before persisting).

Surface the four-step launch → inspect → heal → memorize sequence from the frontmatter composition as an explicit numbered workflow in the body.

DimensionReasoningScore

Conciseness

The body is lean and assumes Claude's competence — a one-line purpose statement plus a fallback — with no padding or explanation of concepts Claude already knows.

5 / 5

Actionability

Guidance is high-level only: 'Manually start a tmux session, scrape output, ask the LLM to diagnose, record the resolution' gives hints but no concrete commands, code, or tmux invocations to execute.

2 / 5

Workflow Clarity

A rough watch → diagnose → record sequence is implied, but steps are poorly defined and validation is absent; because the heal step mutates state (destructive), the missing validation/feedback loop caps this below the midpoint.

2 / 5

Progressive Disclosure

This is a simple sub-50-line skill with no bundle files and no need for external references; its two short sections are well-organized, satisfying the simple-skill exception for progressive disclosure.

5 / 5

Total

14

/

20

Passed

Description

51%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 specific and distinct but reads as an internal deprecation/risk memo rather than a user-facing trigger description: it lacks natural trigger phrases and a positive 'Use when...' clause, capping completeness and trigger-term quality.

Suggestions

Lead with a concise user-facing 'Use when ...' trigger clause (e.g. long-running builds, CI image builds, regression suites that fail intermittently) before the deprecation rationale.

Replace internal plan references ('§3.1 A1/A8', 'E5 bounded sub-agent contract') with plain-language trigger keywords a user would actually say.

Keep the deprecation and risk/capability requirements, but move them after the trigger guidance so the what/when comes first.

DimensionReasoningScore

Specificity

Names the domain and several concrete actions — 'launches arbitrary commands from the user message in tmux and lets sub-agent auto-apply a fix' — with only minor coverage gaps around the inspect/memorize steps.

4 / 5

Completeness

It clearly states what the skill does, but the 'when' is a deprecation/disablement notice rather than a positive 'Use when...' trigger clause, so completeness is capped at 3 per the missing-trigger-guidance rule.

3 / 5

Trigger Term Quality

Beyond 'Build watchdog' there are few natural user-facing trigger terms; the bulk is internal jargon ('E5 bounded sub-agent contract', 'Jinja sandbox', 'saga-style compensation') users would never say.

2 / 5

Distinctiveness Conflict Risk

The tmux + sub-agent build-watchdog niche is fairly distinct and unlikely to fire for unrelated skills, with only minor overlap risk against general monitoring skills.

4 / 5

Total

13

/

20

Passed

Validation

81%

Checks the skill against the spec for correct structure and formatting. All validation checks must pass before discovery and implementation can be scored.

Validation13 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

metadata_field

'metadata' should map string keys to string values

Warning

frontmatter_unknown_keys

Unknown frontmatter key(s) found; consider removing or moving to metadata

Warning

Total

13

/

16

Passed

Repository
TokenRhythm/opensquilla
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.