CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

sce-commit

Analyze staged changes and run the regular or explicit bypass commit workflow

Quality

80%

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 ./.claude/skills/sce-commit/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

88%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 an exceptionally disciplined workflow controller: exact commands, precise input-parsing rules, exhaustive status branching, validation gates at every destructive boundary, and clean delegation of phase, layout, and style content to real reference files. Its only deductions are deliberate rule repetition across sections and one two-level reference chain (via atomic-commit.md to commit-message-style.md).

Suggestions

Consolidate the thrice-repeated instruction to read references/atomic-commit.md (Phase reference, both path steps, Rules) into a single authoritative statement to tighten conciseness.

Mention references/commit-message-style.md's role from SKILL.md (or flatten the chain) so no reference sits two hops from the entry point, improving the one-level-deep structure.

Trim the abstract execution-contract prose in the opening section (e.g., the 'Non-SCE helpers may assist' paragraph) to the minimum needed for correct behavior.

DimensionReasoningScore

Conciseness

The body is tight, imperative control flow with zero explanation of concepts Claude already knows — every line is workflow contract ("Run `git diff --cached --quiet`. A zero exit status means nothing is staged."). It is not a 5 because several guard rules are repeated across sections ("Read `references/atomic-commit.md` before running the phase" appears three times; the never-stage rule appears in the workflow steps, bypass path, and Rules), and the opening execution-contract prose ("Non-SCE helpers may assist...") is abstract padding that could be trimmed. It is well above 3 since there is no unnecessary explanation, only deliberate repetition.

4 / 5

Actionability

Guidance is fully executable: one exact command with exit-status semantics ("Run `git diff --cached --quiet`. A zero exit status means nothing is staged"), precise input parsing ("first whitespace-separated token is exactly `oneshot` or `skip`, compared case-insensitively"), a complete status-to-layout branch map (`blocked`/`proposal`/`bypass_message`), and named layouts resolved to a real file (`references/output.md` — verified present). Per the rubric's scoring note, absence of code in an instruction-only skill is not penalized when the guidance is this actionable; nothing needed to execute a step is left implicit.

5 / 5

Workflow Clarity

Both paths are numbered with explicit gates and validation at every boundary: a user-confirmation staging gate ("Wait for the user's confirmation. Do not stage files on their behalf"), a pre-flight staged-content check with exit-code semantics, blocked-status branching before any commit, defined failure handling ("render the **Bypass Git failure** layout ... and stop"), and even detection of a contract violation ("Treat a `proposal` result as a contract violation: report it and stop without committing"). The destructive-operation cap does not apply since validation/verification steps are present throughout, and the Rules section acts as a checklist; the no-retry policy is an explicit, deliberate terminal failure path rather than a missing checkpoint.

5 / 5

Progressive Disclosure

Scoring against the actual bundle (references/atomic-commit.md, references/output.md, references/commit-message-style.md — all real, verified files): SKILL.md appropriately holds only control flow and clearly signals each reference at its point of use ("Read `references/atomic-commit.md` before running the phase, not after"), with the why explained ("A regular run that stops at the staging gate ... both end without ever needing it"). It is not a 5 because the reference chain is two levels deep from the skill (SKILL.md → atomic-commit.md → commit-message-style.md), which exceeds the rubric's one-level-deep ideal, though every hop is well-signaled and navigation remains easy. It is clearly above 3: content is correctly split across files rather than inlined.

4 / 5

Total

18

/

20

Passed

Description

61%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 concise, concrete, and third-person, cleanly naming the domain (staged changes, commit workflow) with two distinct paths. Its main weakness is the complete absence of any 'when to use this' trigger guidance, and it never mentions the skill's primary output — commit messages — which also limits natural trigger terms like "git" or "commit message".

Suggestions

Append an explicit trigger clause, e.g. "Use when the user asks to commit, write commit messages, or split staged changes into atomic commits" — this lifts completeness from 3 toward 5.

Mention the skill's actual output (generating commit messages / proposing commit splits) in the what-clause to sharpen specificity and add the natural trigger terms "commit message" and "git".

Include the literal mode tokens ("oneshot"/"skip") or the phrase "bypass confirmation" so users invoking the bypass path have a natural phrase that matches.

DimensionReasoningScore

Specificity

The description states two concrete actions — "Analyze staged changes" and "run the regular or explicit bypass commit workflow" — in a clearly named git-commit domain, matching the anchor for naming a domain with 1-2 concrete actions. It is not a 4 because it omits the skill's core output (generating commit messages, proposing splits) and lists no actions beyond analyze/run; it is above a 2 because the actions are domain-specific rather than generic.

3 / 5

Completeness

The "what" is clear ("Analyze staged changes and run the regular or explicit bypass commit workflow"), but there is no "Use when..." clause or equivalent trigger guidance anywhere — matching the anchor 'clear what, when missing', and the judging guideline caps completeness at 3 for a missing 'Use when...' clause. It cannot be 4 without any explicit when; it exceeds 2 because the what is concrete, not vague.

3 / 5

Trigger Term Quality

Natural terms a user would say are present — "staged changes", "commit", "commit workflow", "bypass" — giving good keyword coverage in the commit domain. It falls short of a 5 because common variations are missing: "git", "commit message", and the literal mode tokens (`oneshot`, `skip`) never appear; "explicit bypass" leans toward internal jargon rather than a user phrase.

4 / 5

Distinctiveness Conflict Risk

"staged changes" and "regular or explicit bypass commit workflow" carve a fairly clear niche (staging-based commit workflow) distinct from generic skills, with only minor overlap risk against closely related commit-message skills. It is not a 5 because "commit" is a crowded space where a generic 'generate commit messages' skill could compete for the same trigger.

4 / 5

Total

14

/

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

crocoder-dev/shared-context-engineering

Path

.claude/skills/sce-commit/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.