CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

resolve-repo

Resolve the correct repository for a task — detect VCS platform, match ticket to repo, clone if missing. Run this in a sub-agent to avoid filling main context.

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 ./skills/resolve-repo/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

73%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.

A highly actionable, well-sequenced workflow with excellent error-recovery and validation checkpoints. Its main weakness is redundancy — several constraints are stated three times across the Context Budget, per-step rules, and the Important section — which hurts token efficiency in a skill explicitly designed to be lightweight.

Suggestions

De-duplicate the constraint restatements: state each rule once (e.g., keep "never read repo contents/READMEs" only in the Context Budget or Important section, not in both plus step 3), saving tokens in a skill that must stay lightweight.

Trim or fold the Important section — four of its six bullets restate constraints already given in Context Budget and steps 1-4 — or merge it into Context Budget.

Make the remote-clone step fully concrete by giving the URL construction (e.g., "https://<platform-host>/<group>/<repo>.git") instead of "construct URL from platform + group + repo name".

DimensionReasoningScore

Conciseness

The body is dense and imperative with no explanation of concepts Claude already knows, but constraints are restated repeatedly: "never read repo contents/READMEs" appears in Context Budget, step 3 ("no cloning, no README reading, no file inspection"), and Important; the "ALWAYS emit the Repo Match Results block" rule appears in steps 2, 4, and again in the output rules; the Important section largely duplicates earlier constraints. This is more than anchor 4's "minor instances" of trimming but well short of anchor 2's padded verbosity.

3 / 5

Actionability

Mostly executable guidance: the repo-scan bash loop with jq extraction, concrete glab/gh API commands with jq filters, the determinism-lib sourcing pattern, and the clone/setup-hint checks are copy-paste ready. Minor gaps keep it below anchor 5: "construct URL from platform + group + repo name" and `git clone <remote-url>` are placeholders rather than concrete commands.

4 / 5

Workflow Clarity

Six clearly sequenced steps with explicit conditional branches (a HIGH local match skips remote lookup but never skips the results block; ambiguity routes to step 3), error-recovery fallbacks for every failure mode (lib abstain → AI inference; CLI missing → local-only + install instructions; no repos → ask user), a user-confirmation checkpoint before cloning, machine-checked pinned output formats acting as checklists, and post-clone setup verification. Cloning is user-confirmed, non-batch, and reversible, so the destructive-operation cap does not apply.

5 / 5

Progressive Disclosure

No bundle files exist; the single external reference (~/.claude/lib/determinism/CONTRACT.md) is one level deep and clearly signaled, and the file is well-organized with clear section headers. It sits at anchor 4 rather than 5 because at ~200 lines of dense rules, some content (the pinned output-format specs, determinism contract details) could arguably be externalized, though inlining is defensible for a sub-agent skill.

4 / 5

Total

16

/

20

Passed

Description

66%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 specific, action-oriented description with good natural keywords and a distinct niche. Its main weakness is the missing "Use when..." trigger clause, which caps completeness at 3 and slightly weakens trigger-term coverage and distinctiveness.

Suggestions

Add an explicit "Use when..." clause, e.g., "Use when a task or ticket doesn't name its repository, when the user asks which repo to work on, or before starting work in a multi-repo workspace."

Include natural trigger variations users would say — "which repo", "workspace", "clone the right repository" — to improve trigger-term coverage and auto-triggering.

Consider naming the input context (Jira/Linear/GitHub ticket) in the trigger clause to further distinguish it from generic git-clone or repo-setup skills.

DimensionReasoningScore

Specificity

The description lists several concrete actions — "detect VCS platform, match ticket to repo, clone if missing" — under the umbrella task "Resolve the correct repository for a task". It falls short of anchor 5 because coverage has minor gaps (local-first scanning, scoped remote lookup) but exceeds anchor 3's 1-2 actions.

4 / 5

Completeness

The "what" is clear (detect platform, match ticket to repo, clone if missing), but there is no "Use when..." clause or equivalent explicit trigger guidance — "Run this in a sub-agent" is a how, not a when. Per the judging guideline, a missing "Use when..." clause caps completeness at 3.

3 / 5

Trigger Term Quality

Natural keywords users would say are present: "repository", "repo", "ticket", "clone", "VCS platform". A few natural trigger phrases are missing (e.g., "which repo", "workspace", "where should this change go"), so it matches anchor 4 rather than anchor 5's comprehensive synonym coverage, but is stronger than anchor 3's partial coverage.

4 / 5

Distinctiveness Conflict Risk

It occupies a clear niche (ticket-to-repo resolution) with distinct terms (ticket, repo, clone, VCS platform). Minor overlap risk remains with generic git/clone or repo-setup skills because no trigger scenario disambiguates, keeping it at anchor 4 rather than 5.

4 / 5

Total

15

/

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

AndreJorgeLopes/devflow

Path

skills/resolve-repo/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.