CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

cw-dogfood

Use when a Codewhale change needs proving in the real product, or when asked to build/install/dogfood the local binaries: stamped release build, atomic install, fresh-shell verification, and the manual QA that gates cannot cover.

Quality

92%

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

SKILL.md
Quality
Evals
Security

Quality

Content

96%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: fully executable commands, a gated workflow with real validation checkpoints, and lean product-specific detail with zero padding. The only structural improvement is moving the detailed QA scenario scripts into a referenced file so SKILL.md stays a tighter overview.

DimensionReasoningScore

Conciseness

The body is imperative and dense with no explanations of concepts Claude already knows; even the 'why' passages are project-specific hazards Claude couldn't know ('poisons the kernel's code-signature cache for the inode, and later execs hang until reboot'). Every token carries an operating instruction or a real constraint.

5 / 5

Actionability

Copy-paste-ready commands throughout: the stamped build invocation with CODEWHALE_BUILD_SHA, 'scripts/release/install-dogfood.sh', the fresh-shell check "zsh -lc 'type -a codew codewhale; codew --version'", and the app-server smoke commands. The only placeholder (--model <model>) is inherently variable and justified.

5 / 5

Workflow Clarity

Seven numbered steps with explicit validation checkpoints: 'Never install an ungated build', the installer refusing unstamped/dirty trees, and the fresh-shell SHA check guarding against the 'classic false pass'. Error-recovery paths are present (deliberate CODEWHALE_ALLOW_DIRTY_DOGFOOD=1 override with disclosure, 'attach logs if the freeze reproduces'), plus a red-flags checklist.

5 / 5

Progressive Disclosure

Well-organized sections (When to use / Workflow / Red flags / Output) with clearly signaled one-level-deep links (docs/RELEASE_CHECKLIST.md, crates/tui/AGENTS.md, docs/MOTION_CONTRACT.md, sibling skills). The inline scenario scripts (e.g. the issue-numbered #3289 Windows Terminal replay procedure) are heavy inline content that could be split into a reference file, which keeps this below a 5.

4 / 5

Total

19

/

20

Passed

Description

87%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 strong description with an explicit 'Use when' trigger clause, a concrete action list, and a clearly distinct niche within the Codewhale skill loop. It could be marginally sharper by surfacing the additional natural trigger phrases already present in the body and making the 'proving' framing more concrete.

Suggestions

Fold in the natural trigger phrases the body already enumerates (e.g. 'install the build', 'get this on my machine', 'release candidate QA') so users phrasing the request in those words match the skill.

Replace or supplement the abstract 'needs proving in the real product' with a concrete capability verb (e.g. 'exercise the installed binary in a real terminal') to sharpen the 'what'.

DimensionReasoningScore

Specificity

Names several concrete actions — 'stamped release build, atomic install, fresh-shell verification, and the manual QA that gates cannot cover' — but the framing 'a change needs proving in the real product' is somewhat abstract, so coverage has minor gaps rather than being fully comprehensive.

4 / 5

Completeness

Explicitly answers both questions: 'Use when a Codewhale change needs proving in the real product, or when asked to build/install/dogfood the local binaries' gives concrete triggers, and the colon-list of operations gives a clear 'what'.

5 / 5

Trigger Term Quality

Contains genuine natural phrasing users would say ('build/install/dogfood the local binaries', 'proving in the real product'), and 'dogfood' is exactly the word a user would use. A few natural variants the body itself lists ('install the build', 'get this on my machine', 'QA') are absent from the description.

4 / 5

Distinctiveness Conflict Risk

Scoped to the named product and a distinct stage of a six-skill loop; 'dogfood' and 'install the local binaries' are triggers unique to this skill and do not overlap with the gates/land/handoff siblings.

5 / 5

Total

18

/

20

Passed

Validation

87%

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

Validation — 14 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

relative_links

Relative link issues: 6 suspicious

Warning

referenced_paths_exist

Referenced path issues: 3 missing, 2 deeper-than-1-level

Warning

Total

14

/

16

Passed

Repository

Hmbown/Codewhale

Path

docs/skills/cw-dogfood/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.