CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

ce-babysit-pr

Babysits an open GitHub PR until merge-ready. Use when asked to watch a PR over time — not for one-shot comment resolution or one CI failure. GitHub (incl. Enterprise) only.

Quality

89%

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

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.

An exceptionally disciplined orchestration skill: concrete commands and machine-checkable conditions throughout, a fully sequenced tick with validation and error-recovery loops, and textbook progressive disclosure — a lean overview whose every reference target exists and is signaled at point of use. The only cost of its compression is minor rule duplication across the boundaries and steps, which keeps conciseness just short of the top anchor.

DimensionReasoningScore

Conciseness

The body is extremely dense with project-specific, non-obvious rules and explains nothing Claude already knows (no gh/GitHub/CI basics), assuming full competence. It sits between anchors 4 and 5: nearly every token encodes a real rule, but the snapshot-is-the-only-authority principle is stated three times across the overview and boundaries (lines 13, 26, 27) and a few boundary rules are restated in the steps, so minor consolidation is possible — fitting anchor 4 (efficient, minor trimming possible) more than anchor 5's 'every token earns its place'.

4 / 5

Actionability

Guidance is concrete and copy-paste-ready: exact commands (`gh pr checkout <ref>`, `gh run rerun <run-id> --failed -R <host>/<owner>/<repo>`, `gh stack merge` + `gh stack sync`), exact delegate invocations (`ce-resolve-pr-feedback mode:pipeline`, `ce-debug mode:pipeline`), and exact machine-readable conditions (`manager_status == "confirmed"`, `open_needs_human == 0`, `branch_currency_blocker == null`). This matches anchor 5 for an instruction-only skill: fully executable commands covering the common cases, far above anchor 4's 'minor gaps'.

5 / 5

Workflow Clarity

The four-step process is explicitly sequenced with an ordering invariant inside each tick, and validation checkpoints are exhaustive: every arming step has a stop path (repo check, PR resolution, checkout/push access), the tick opens with a terminal check, stale-SHA cancellation invalidates dead CI, and defect states (`unrequested_base_merge`, red residuals) are reported rather than silently dropped. Feedback loops for error recovery (flaky → rerun, real failure → ce-debug, unfixed → tracked residuals) match anchor 5; the destructive-operation validation cap is satisfied by the atomic claim, `expected_head_sha`, and defect-reporting steps.

5 / 5

Progressive Disclosure

Scored against the actual bundle: every `references/*.md` path cited in the body exists on disk, the single script (`scripts/pr-snapshot`) is real, and references are one level deep (`stack.md`/`watch-loop.md` reach `stack-commands.md`, no deeper nesting). The body is a lean ~60-line overview that signals each per-topic file exactly at its point of use (branch-currency.md at the currency step, settle.md at stop conditions, tick.md before first snapshot), matching anchor 5's clear overview with well-signaled one-level-deep references and easy navigation.

5 / 5

Total

19

/

20

Passed

Description

82%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 tight, third-person description that explicitly answers both what and when, with well-chosen negative triggers and platform boundaries that carve out a distinct niche. Its only weakness is that the single composite action leaves the concrete capability streams unlisted, capping specificity at the middle anchor.

Suggestions

Enumerate the concrete action streams in the description (e.g. 'resolves review comments, reruns flaky CI, and handles branch updates') to lift specificity from one composite action to several listed ones.

Add one or two natural synonyms users might say, such as 'monitor' or 'pull request' spelled out, to broaden trigger-term coverage.

DimensionReasoningScore

Specificity

The description names one concrete composite action — "Babysits an open GitHub PR until merge-ready" — with a clear outcome, but does not enumerate the specific sub-actions the skill performs (resolving review comments, rerunning failed CI, handling branch updates). This matches anchor 3 (domain plus 1-2 concrete actions, not comprehensive) better than anchor 4, which requires several listed actions, and better than anchor 2, since 'babysits... until merge-ready' is more concrete than a generic domain mention like 'Processes PDF files'.

3 / 5

Completeness

Both halves are explicit: what — "Babysits an open GitHub PR until merge-ready" — and when — "Use when asked to watch a PR over time — not for one-shot comment resolution or one CI failure", plus the platform boundary "GitHub (incl. Enterprise) only". This clearly matches anchor 5 (explicit what AND when with concrete trigger phrases); anchor 4 would require a weaker or less explicit 'when' clause.

5 / 5

Trigger Term Quality

Natural trigger phrases are present — "watch a PR over time", "GitHub PR", "merge-ready" — and the explicit negative triggers ("not for one-shot comment resolution or one CI failure") sharpen matching. A few natural synonyms users would say are missing (e.g. "pull request" spelled out, "monitor", "keep an eye on"), which fits anchor 4 (good coverage, a few natural terms missing) rather than anchor 5's comprehensive synonym coverage.

4 / 5

Distinctiveness Conflict Risk

A clear niche (ongoing PR supervision) with explicit exclusion of adjacent one-shot tasks ("not for one-shot comment resolution or one CI failure") and a platform boundary ("GitHub only"), minimizing overlap with resolve/debug/merge skills. This matches anchor 5's clear niche with distinct triggers and minimal conflict risk.

5 / 5

Total

17

/

20

Passed

Validation

93%

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

Validation — 15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

frontmatter_unknown_keys

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

Warning

Total

15

/

16

Passed

Repository

EveryInc/compound-engineering-plugin

Path

skills/ce-babysit-pr/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.