CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

rebase

Start a local Git rebase when the user explicitly requests replay of a named source ref, or of the current branch, onto a named target, including a rebase-then-push request, which carries exact-lease publication to the branch's own remote destination, or continue or abort an active rebase. Use only when local history replay will start or is active; do not use for merge-based branch updates, repository merge settings, pull-request or merge-request merging, or push-only requests.

66

Quality

83%

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

71%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 achieves exemplary progressive disclosure and leaves no validation gap in a high-stakes destructive workflow, at the cost of near-total command delegation to references and an extremely dense mermaid state machine that trades readability for completeness. Conciseness is good overall but marred by repeated boilerplate recovery edges.

Suggestions

Add a short numbered 'happy path' sequence for the common case (start → fetch/reconcile → replay → resolve conflicts → verify → publish), keeping the full mermaid state machine in SKILL.md or a reference for edge cases — this would raise both actionability and workflow clarity without adding much length.

Inline the one or two core commands (the 'git rebase --onto' invocation and the 'git push --force-with-lease=<ref>:<oid>' form from publication.md) directly in SKILL.md so the most common path is executable without opening a reference.

State the 'Failure or unobservable → Recover' rule once as a global convention instead of repeating it on ~15 chart edges, trimming the mermaid chart's token cost and visual noise.

DimensionReasoningScore

Conciseness

There is no padding, no tutorial content, and no re-explanation of Git concepts Claude already knows — the prose is maximally compressed ('Preserve unrelated work; a passing dirty worktree does not validate its commit'). This clears the 'efficient; minor instances of over-explanation' anchor rather than the lean 5, because the ~100-line mermaid chart repeats boilerplate guard edges ('Failure or unobservable → Recover' appears ~15 times) that could be stated once as a global rule.

4 / 5

Actionability

The body supplies concrete decision-level guidance — exact conditions, dispositions, and checkpoints like 'create start-only branch worktree', 'save and bind exact entry', 'push immutable R with the authorized exact lease' — but contains no executable commands; the actual commands (e.g. the git push --force-with-lease invocation) live only in reference files. This matches the 'some concrete guidance but incomplete; missing key details' anchor: the body instructs what to do at each branch but the reader must open a reference for how, and several directives ('bind metadata-owning worktree/Git dir') are unexplained in place.

3 / 5

Workflow Clarity

The sequence is fully specified with pervasive validation and feedback loops — result-bound verification on an isolated checkout, reobservation before completion, correction-then-revalidate loops, and a capped retry loop for lease rejection — so the destructive-operation cap (missing validation) does not apply. It does not reach 5 because the ~50-node mermaid state machine with telegraphic edge labels ('Metadata/unmerged absent; R bound; result ref and bound goal relation match') is genuinely hard to follow, unlike the crisp stepwise sequences in the top anchors.

4 / 5

Progressive Disclosure

The body is an overview that signals five one-level-deep references exactly at the decision points where they are needed ('read [publication](./references/publication.md)', '[named stash](./references/named-stash.md)', etc.); all five referenced files exist, are substantive, and contain no nested references. Content is appropriately split — the decision topology in SKILL.md, executable detail in references — with easy inline navigation, matching the top anchor.

5 / 5

Total

16

/

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 explicit what/when structure, concrete lifecycle actions, and effective negative boundaries that sharply distinguish it from merge-based workflows. Main weakness is that a few natural trigger phrasings and the conflict-resolution capability sit in the body's Keywords line instead of the description.

DimensionReasoningScore

Specificity

The description lists several concrete actions — 'Start a local Git rebase', 'replay of a named source ref, or of the current branch, onto a named target', 'continue or abort an active rebase', and 'exact-lease publication to the branch's own remote destination' — matching the 'several specific actions; minor gaps in coverage' anchor. It is not a 5 because conflict resolution and recovery actions are absent from the description (they only appear in body keywords), and the actions are embedded in qualifying clauses that slightly obscure them.

4 / 5

Completeness

It clearly answers what ('Start a local Git rebase... continue or abort an active rebase') and explicitly answers when with a concrete trigger clause: 'Use only when local history replay will start or is active; do not use for merge-based branch updates, repository merge settings, pull-request or merge-request merging, or push-only requests.' Both what and when are explicit with concrete trigger phrases and negative boundaries, matching the top anchor exactly.

5 / 5

Trigger Term Quality

Natural trigger phrases are present and repeated in user-plausible forms: 'Git rebase', 'rebase-then-push request', 'continue or abort an active rebase'. This fits the 'good keyword coverage; a few natural terms missing' anchor; it falls short of 5 because common phrasings like 'rebase onto', 'interactive rebase', or 'rewrite history' are absent, and the trigger-rich keyword variants ('rebase conflict', 'rewritten history') live in the body's Keywords line rather than the description.

4 / 5

Distinctiveness Conflict Risk

It carves a clear niche — explicit local history replay — and actively de-conflicts from adjacent skills via 'do not use for merge-based branch updates, repository merge settings, pull-request or merge-request merging, or push-only requests'. Minimal overlap risk with merge/pull/push skills, matching the 'clear niche with distinct triggers; minimal conflict risk' anchor.

5 / 5

Total

18

/

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

Jamie-BitFlight/claude_skills

Path

.agents/skills/rebase/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.