CtrlK
BlogDocsLog inGet started
Tessl Logo

loom

Designing a repository's operating layer — which project-local skills, recipes, and orchestration workflows it needs, as one system. Not for a single skill body (Sigil) or runtime routing (Nexus).

50

Quality

54%

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 ./.archive/loom/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

55%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 skill has excellent workflow discipline — a clear phased chain with real validation gates and per-recipe VERIFY criteria — but it ships without any of the reference files its guidance depends on, and the body duplicates its own chain, handoffs, and thresholds across four sections. As delivered it is a well-designed index to a bundle that is not there.

Suggestions

Ship the referenced bundle files (reference/blueprint-method.md, recipe-design.md, workflow-design.md, delegation-handoffs.md, validation-checklist.md, autorun-schema.md, and the _common/*.md files) or inline their essential content — currently every 'Read When' pointer is dangling.

Consolidate the phase-chain documentation: keep the canonical Workflow table and have the Recipes, Operating Flows, and Subcommand Dispatch sections reference it instead of restating it, removing roughly a third of the body.

Collapse the triple-duplicated collaboration material (HTML comment block, ASCII diagram, and handoff table) into the single Collaboration table and move per-agent payload details to reference/delegation-handoffs.md.

DimensionReasoningScore

Conciseness

The body is noticeably padded through duplication: the SURVEY→MAP→DESIGN→DELEGATE→VERIFY chain appears four times (Workflow table, Operating Flows, Recipes table, Subcommand Dispatch), the delegation handoffs are listed three times (HTML comment block, ASCII collaboration diagram, Collaboration table), and the Critical Thresholds table restates Core Contract items. It avoids explaining concepts Claude already knows, which keeps it above level 1, but the redundancy is well beyond 'minor instances that could be trimmed'.

2 / 5

Actionability

There is genuinely concrete guidance — exact thresholds ('≤ 5 sequential phases', 'overlap under 30%', '≤ 7 specialists per orchestrator'), a dispatch rule for the first input token, and a copyable NEXUS_HANDOFF schema — but the actionable detail for each phase (blueprint method, recipe patterns, topology selection, delegation payloads, the 14-item checklist) lives entirely in `reference/*.md` and `_common/*.md` files that are absent from the bundle, leaving key steps unexecutable as shipped. That matches 'some concrete guidance but incomplete; missing key details' rather than level 4.

3 / 5

Workflow Clarity

The phase chain is explicit and consistently applied, each phase has a stated purpose and a named reference to load on entry, each recipe has its own **VERIFY** gate in addition to the universal discipline, and the 14-item checklist blocks delivery until it passes — a textbook validation checkpoint with feedback-loop semantics ('Delivery is blocked until it passes').

5 / 5

Progressive Disclosure

The structure is well designed — a Reference Map table with 'Read This When' conditions per file, references one level deep, clearly signaled per phase — but none of the referenced files (7 `reference/*.md` and 4 `_common/*.md`) exist in the bundle, so navigation leads nowhere, and substantial content that belongs in those files (per-recipe behavior notes, the collaboration handoff tables, threshold rationale) is inlined in the ~295-line body instead. This sits between 'references present but not clearly signaled' (they are signaled) and 'most content appropriately placed' (it is not, given the missing bundle), so 3 rather than 4.

3 / 5

Total

13

/

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 defines a crisp, well-scoped niche with useful negative triggers, but it is missing any positive 'use when' guidance and lists components rather than actions, so a user wanting this skill has no natural trigger phrases to say. It reads as a boundary statement more than an invocation contract.

Suggestions

Add an explicit 'Use when...' clause with concrete trigger phrases, e.g. 'Use when designing a coordinated set of project-local skills, recipes, or workflows for a repository, or when auditing an existing .claude/ layer for gaps and overlap.'

Replace the component list with 2-3 action verbs (e.g. 'Blueprint, map, and audit a repository's operating layer — skills, recipes, workflows, and the routing map') to raise specificity.

Include natural synonyms users would actually say ('project skill suite', 'agent routing map', 'task playbooks', 'repo automation layer') to broaden trigger-term coverage.

DimensionReasoningScore

Specificity

The description names the domain ('a repository's operating layer') and its concrete components ('project-local skills, recipes, and orchestration workflows'), but offers only one action verb — 'Designing' — so it matches the anchor 'names domain and 1-2 concrete actions, but not comprehensive' rather than the level-4 anchor's list of several distinct actions.

3 / 5

Completeness

The 'what' is clear and specific, but there is no 'Use when...' clause or equivalent positive trigger guidance — only negative scoping ('Not for a single skill body (Sigil) or runtime routing (Nexus)'), which the rubric explicitly caps at 3. It is not level 2 because the 'what' is concrete, not vague.

3 / 5

Trigger Term Quality

Relevant keywords are present ('operating layer', 'skills', 'recipes', 'orchestration workflows', 'repository') but natural user phrasings for the task — e.g. 'design our project skills', 'agent suite', 'routing map', 'task playbook' — are missing, matching the 'some relevant keywords but missing common variations or synonyms' anchor rather than level 4's 'good keyword coverage'.

3 / 5

Distinctiveness Conflict Risk

The explicit exclusions ('Not for a single skill body (Sigil) or runtime routing (Nexus)') carve a clear niche with low conflict risk against the named siblings. It falls short of level 5 because other adjacent design roles (e.g. a global/eco-system agent designer) are not excluded and there are no positive distinct trigger phrases.

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
simota/agent-skills
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.