CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

loki-seal

Use when finishing any coding task in a repo with tests. Loki Seal runs the repo's real test suite and refuses to let you declare done while tests newly fail, or while tests or CI config were deleted, skipped, xfailed or weakened. Runs locally with no model calls.

Quality

85%

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

75%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 lean, information-dense body that explains only what the hook enforces and gives concrete guidance for being blocked and for installation. Its weaknesses are structural rather than substantive: item 3's contract rules are packed into one unreadable sentence, the recovery loop after a block is implicit, and the README/hook-file references point outside the bundle without a clearly signaled reference location.

Suggestions

Break item 3's contract rules into a short bulleted list (or move the sentence-classification details into a reference file) instead of one run-on sentence — this would tighten conciseness and improve actionability of the most complex rule.

Add an explicit feedback step under 'What to do when it blocks', e.g. 'after fixing, attempt to stop again — the hook re-runs the suite and re-checks the contract', to close the implicit recovery loop.

Either bundle the README (contract spec) and note bin/loki-seal.js as a plugin file in a clearly labeled references section, or drop the '(see README)' inline pointer since no such file exists in the skill bundle.

DimensionReasoningScore

Conciseness

The body is dense with non-obvious rules and explains no known concepts, but item 3 packs the full contract-parsing and sentence-classification rules into a single run-on sentence, with minor redundancy ('always blocks while any test matched to a request item fails now: failing tests never count as covered'). Not 5 because that over-packed detail could be trimmed or restructured; not 3 because there is no padding or over-explanation.

4 / 5

Actionability

Concrete executable guidance appears in 'Fix the production code. Do not delete, skip, xfail or weaken tests' and the install command 'claude plugin marketplace add <owner/repo or path>'. Not 5 because much of the body describes the hook's behavior rather than instructing the reader, and the 'what to do' section is brief; not 3 because the guidance given is specific and executable, not vague.

4 / 5

Workflow Clarity

The hook's behavior is a clear numbered sequence (detect runner, block conditions, contract check, receipt) and the blocked-response path includes an escalation branch ('If a test is genuinely wrong, tell the user'). Not 5 because the recovery loop is implicit — it never says to re-attempt finishing or that the hook re-runs the suite — leaving a minor validation/feedback gap.

4 / 5

Progressive Disclosure

The body is well-organized into short sections (hook behavior, what to do when blocked, install) and is under 50 lines, but it cites '(see README)' and 'bin/loki-seal.js' inline — files not present in this skill bundle (no references/ directory exists) — so those pointers are not clearly signaled or verifiable. Not 5 because of that dangling reference; not 3 because structure is otherwise good and content is appropriately placed.

4 / 5

Total

16

/

20

Passed

Description

88%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: it explicitly states when to trigger, what the skill does in concrete third-person terms, and its enforcement niche. Keyword coverage is good but a few natural trigger synonyms are missing, and the broad 'any coding task' trigger introduces minor overlap risk with related testing skills.

DimensionReasoningScore

Specificity

Multiple concrete actions are named: 'runs the repo's real test suite', 'refuses to let you declare done while tests newly fail', 'deleted, skipped, xfailed or weakened', and 'Runs locally with no model calls' — comprehensive coverage of what the skill enforces. Not 4 because no capability area is left uncovered for this skill's scope.

5 / 5

Completeness

Explicitly answers both: 'what' — runs the real test suite and blocks completion on new failures or weakened/removed tests; 'when' — 'Use when finishing any coding task in a repo with tests', a concrete trigger phrase. Voice is third person ('Loki Seal runs...'), matching the top anchor.

5 / 5

Trigger Term Quality

Natural phrases users would say are present ('finishing any coding task', 'repo with tests', 'test suite', 'CI'), but common synonyms like 'green', 'verify before done', or 'run the tests before finishing' are missing. Not 3 because coverage is good and clearly matches how users phrase completion requests; not 5 because a few natural variations are absent.

4 / 5

Distinctiveness Conflict Risk

Clear niche (test-integrity gating at task completion) with distinct triggers around 'declare done' and test weakening, but the trigger 'any coding task in a repo with tests' is broad and could overlap with general testing/verification skills. Not 5 because of that minor overlap risk with closely related test-running skills.

4 / 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

asklokesh/loki-mode

Path

packages/loki-seal/skills/loki-seal/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.