CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

sce-commit

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

Quality

76%

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 ./.opencode/skills/sce-commit/SKILL.md

The canonical home for this skill is sce-commit in crocoder-dev/shared-context-engineering

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.

This is a well-engineered dispatcher document: precise input parsing, two unambiguous paths, validation gates and failure handling at every risky step, and disciplined delegation of phase and layout detail to reference files. The only deductions are repeated guardrail statements across sections and one bundle reference (commit-message-style.md) that is nested two levels deep rather than surfaced from the top level.

Suggestions

Consolidate the repeated prohibitions (do-not-stage, do-not-commit) so each appears once — either in the workflow steps or in the Rules section — and trim the generic 'Execution contract' preamble.

Surface `references/commit-message-style.md` from SKILL.md (e.g., one line in the Phase reference section noting the phase writes messages in that style) so no reference file is two levels deep.

DimensionReasoningScore

Conciseness

The body is lean and imperative throughout — no explanations of git concepts Claude already knows, and every instruction is operational. It falls short of 5 only because guardrails are repeated across sections ('do not stage' appears in Input, both workflow paths, and Rules again; 'Do not run git commit' is stated twice) and the generic 10-line 'Execution contract' preamble could be trimmed.

4 / 5

Actionability

Guidance is fully executable: an exact command with exit-code semantics ('Run `git diff --cached --quiet`. A zero exit status means nothing is staged.'), precise mode-token parsing rules (exact case-insensitive first-token match, everything else is commit context), explicit branch conditions on named statuses (blocked/proposal/bypass_message), and named output layouts per terminal state. An agent can execute the entire workflow with no invented steps; the reference files carry the remaining concrete commands.

5 / 5

Workflow Clarity

Both paths are strictly sequenced with numbered steps, explicit stops, and validation checkpoints at every risky point: a user-confirmation staging gate before the phase, a staged-content check before any bypass work, status-based branching with a defined contract-violation handler (proposal in bypass mode), and distinct failure/success layouts after the commit handoff. The destructive operation (git commit) is wrapped in validation and explicit failure handling, so the missing-validation cap does not apply.

5 / 5

Progressive Disclosure

SKILL.md is a clean control-flow overview that appropriately delegates the phase (atomic-commit.md) and all user-visible layouts (output.md) to well-signaled, one-level references, each referenced by exact path and read-timing instruction. It falls short of 5 because the third bundle file, references/commit-message-style.md, is never surfaced from SKILL.md — it is only reachable through atomic-commit.md, making that chain two levels deep, which the 5 anchor's one-level-deep criterion excludes.

4 / 5

Total

18

/

20

Passed

Description

53%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, third-person, and accurately states what the skill does, but it omits any 'when to use' trigger guidance and lacks the natural keyword variations (commit message, git commit, oneshot/skip) that would let a user reliably invoke it. It reads as a partial capability statement rather than a complete dispatch description.

Suggestions

Append an explicit trigger clause, e.g. 'Use when the user asks to commit staged changes, wants atomic commit messages, or requests a single oneshot/skip commit.'

Include natural synonyms users would say — 'commit message', 'git commit', 'atomic commits' — alongside 'staged changes'.

Briefly state the two outcomes (regular path proposes commits without committing; bypass path creates exactly one commit) so the 'what' is comprehensive, not just two actions.

DimensionReasoningScore

Specificity

The description names the domain ("staged changes", "commit workflow") and two concrete actions ("Analyze staged changes", "run the regular or explicit bypass commit workflow"), but stops there — it never mentions writing or splitting commit messages, proposing commits, or the one-commit-per-invocation behavior. This matches the anchor for naming a domain with 1-2 concrete actions without comprehensive coverage; a 4 would require several specific actions with only minor gaps.

3 / 5

Completeness

The 'what' is clear (analyze staged changes, run a regular or bypass commit workflow), but there is no 'Use when...' clause or any equivalent trigger guidance, which caps completeness at 3 per the judging guidelines. A 4 would require an explicit — even if imperfect — 'when' statement.

3 / 5

Trigger Term Quality

It contains a few relevant natural terms ("commit", "staged changes", "bypass"), but misses common variations users would actually say — "commit message", "git commit", "atomic commits", the `oneshot`/`skip` tokens, or any extension cues. Some relevant keywords but missing common variations and synonyms places it at the middle anchor rather than 4's good keyword coverage.

3 / 5

Distinctiveness Conflict Risk

The capability described is a specific niche (staged-changes-only atomic commit workflow with a distinct regular/bypass split), giving it minor overlap risk only with closely related generic commit helpers. It falls short of a 5 because the absence of any trigger phrases means a generic 'commit my changes' request has no distinguishing signal pointing to this skill over a sibling commit skill.

4 / 5

Total

13

/

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

.opencode/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.