CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

jujutsu

**REQUIRED** - Always activate FIRST on any git/VCS operations (commit, status, branch, push, etc.), especially when HEAD is detached. If `.jj/` exists -> this is a Jujutsu (jj) repo - raw git commands can corrupt data. Essential git safety instructions inside. DO NOT IGNORE.

1.59x
Quality

78%

Does it follow best practices?

Impact

94%

1.59x

2 of 3 eval scenarios. Add 1 more for a full score.

SecuritybySnyk

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./jujutsu/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

77%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 a highly actionable, safety-conscious jj guide: commands are concrete and executable, workflows have explicit verification and undo loops, and agent-environment pitfalls (pagers, editors, interactive flags) are consistently called out. Its weaknesses are structural and editorial — a long single file that could split advanced topics into reference files, plus repeated reminders and one malformed example block that cost token efficiency.

Suggestions

Move the Workspaces deep-dive and colocated-git switching sections into a references/ file (e.g. WORKSPACES.md, GIT-COLOCATED.md), keeping a short summary plus a clearly signaled link in SKILL.md.

State the `--no-pager` / inline `-m` rules once in the agent-environment section and drop the repeated inline reminders and 'NOT:' comments.

Replace the non-command 'Examples:' code block under 'Creating Atomic Commits' with plain example messages (the same list already appears correctly under 'Preserving Commit Quality').

DimensionReasoningScore

Conciseness

Mostly efficient, command-first content covering genuinely non-obvious jj semantics (working-copy-as-commit, change IDs, fileset quoting), but there is noticeable tightening to do: the `--no-pager` and `-m` guidance is repeated across the intro, individual commands, and the Quick Reference; the 'Creating Atomic Commits' section contains a pseudo-code block of 'Examples:' that is not a command; and the ~35-line Workspaces deep-dive is inlined for a self-described 'rarely-needed feature'. This fits anchor 3 (mostly efficient, some unnecessary explanation or could be tightened) better than anchor 4, where such instances would be minor.

3 / 5

Actionability

Nearly every section gives copy-paste-ready commands with concrete arguments: `jj desc -m "Add user authentication to login endpoint"`, `jj split -m "Add search" src/search.txt`, `jj revert -r <change-id> -B @`, `jj restore 'file:"docs/notes (draft).md"'`, plus a Quick Reference table covering the common cases. It also flags non-executable variants to avoid (`squash -i`, bare `split`), matching anchor 5's fully executable, common-case-covering standard.

5 / 5

Workflow Clarity

Multi-step workflows are clearly sequenced with explicit validation and recovery: 'Describe First, Then Code' (jj st → jj new → jj desc → edit → jj st), 'Verify operations with `jj st` after mutations', a pre-push checklist with bookmark-advance gotchas and `jj undo` recovery, and a 'Preserving Commit Quality' checklist. Destructive operations (squash, abandon, rebase) all carry verification or undo feedback loops, matching anchor 5's explicit validation, error-recovery loops, and checklists.

5 / 5

Progressive Disclosure

Section headers and the Quick Reference give good in-file navigation, and the Workspaces section appropriately defers depth to the official docs, but this ~460-line skill inlines substantial content that would fit reference files (Workspace semantics, colocated git switching, bookmark push workflows) with no bundle files at all. Anchor 3 (some structure, content that should be separate is inline) fits; it is above anchor 2 because structure is not minimal, and below anchor 4 because no well-signaled one-level-deep references exist.

3 / 5

Total

16

/

20

Passed

Description

70%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 a strong safety-first trigger with excellent, explicit activation conditions and natural keywords, but it is lopsided: it says very clearly when to activate and almost nothing about what the skill provides. Adding a brief capability statement (e.g., jj commands for describing, squashing, rebasing, and pushing commits) would raise both completeness and specificity.

Suggestions

Add a one-clause capability statement (e.g., 'Provides jj commands for describing, squashing, splitting, rebasing, and pushing commits') so the 'what' is explicit, not just implied by 'safety instructions inside'.

Include a few more natural VCS trigger terms users would say — merge, rebase, clone, diff — alongside commit/status/branch/push.

Consider softening 'Always activate FIRST on any git/VCS operations' to activate when `.jj/` exists or jj is mentioned, reducing false triggers on plain-git repos.

DimensionReasoningScore

Specificity

The description names the domain ('Jujutsu (jj) repo') and concrete trigger operations ('commit, status, branch, push, etc.') plus a specific detection mechanism ('If `.jj/` exists'), but it never states what the skill itself does beyond 'Essential git safety instructions inside' — the skill's actual capabilities (describe, squash, rebase, bookmarks, push workflows) are absent. This matches anchor 3 (domain plus 1-2 concrete actions, not comprehensive); it is above anchor 2 because the actions and detection details are specific rather than generic, and below anchor 4 because no list of the skill's own actions is provided.

3 / 5

Completeness

The 'when' is explicit and concrete ('Always activate FIRST on any git/VCS operations... especially when HEAD is detached. If `.jj/` exists'), but the 'what' is only weakly stated — 'Essential git safety instructions inside' gestures at content without describing capabilities. This sits between anchors 4 and 5: stronger trigger guidance than anchor 4's example, but the 'what' is too vague for anchor 5's 'clearly and explicitly answers both'.

4 / 5

Trigger Term Quality

Good natural keyword coverage: 'git/VCS operations (commit, status, branch, push)', 'HEAD is detached', 'Jujutsu (jj)', and the `.jj/` marker give users multiple natural phrasings and synonyms. It falls short of anchor 5 because common VCS verbs users would say (merge, rebase, clone, diff, log) are not included.

4 / 5

Distinctiveness Conflict Risk

The `.jj/` existence check carves out a clear niche and disambiguates from plain-git skills, but the instruction to 'Always activate FIRST on any git/VCS operations' fires on every git command, creating minor overlap risk with closely related git skills. Mostly distinct with minor overlap risk — anchor 4 — rather than anchor 5's minimal conflict risk.

4 / 5

Total

15

/

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

allowed_tools_field

'allowed-tools' contains unusual tool name(s)

Warning

Total

15

/

16

Passed

Repository

danverbraganza/jujutsu-skill

Path

jujutsu/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.