conformance-loop

Iteratively improve Fallow analysis accuracy by comparing it with competing tools and verified source truth across a stable real-world corpus.

Quality

61%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

Passed

No findings from the security scan

Fix and improve this skill with Tessl

tessl review fix ./.agents/skills/conformance-loop/SKILL.md

Quality

Content

72%

Reviews the quality of instructions and guidance provided to agents. Good implementation is clear, handles edge cases, and produces reliable results.

The body is admirably lean and well-structured with a clear sequence, but a couple of steps stay abstract and the validation checkpoints are implicit rather than explicit, holding actionability and workflow clarity at the mid anchor.

Suggestions

Make vague steps actionable: for step 1 specify what defines the corpus (e.g. 'list the N projects and the single capability under test'), and for step 3 give the verification method (e.g. 'open the source location Fallow flagged and confirm whether the finding holds').

Frame the regression gate explicitly as a validate/retry loop, e.g. '6. Re-run the full corpus; if the correction does not yield a net improvement, discard it and retry from step 5.'

Add one concrete example of a 'general correction' and its regression fixture so step 5 is copy-paste-ready rather than conceptual.

Dimension	Reasoning	Score
Conciseness	The body is ~10 lines with zero padding: a numbered workflow plus one guardrail ('Competitor output is a lead, not ground truth.'), assuming Claude's competence and explaining no background concepts, matching the 'lean and efficient; every token earns its place' anchor.	3 / 3
Actionability	Some steps are concrete (the TP/FP/FN/model-difference taxonomy, 'regression fixture', and the literal `review` command), but others are vague ('Define the capability and stable project corpus', 'Manually verify disagreements against source') with no concrete commands or detail, fitting 'some concrete guidance but incomplete' rather than fully copy-paste-ready level 3.	2 / 3
Workflow Clarity	The seven steps are clearly sequenced and include verification (step 3) and a net-improvement gate (step 6), but these checkpoints are not framed as explicit validate/fix/retry loops with labeled validation, so checkpoints are implicit rather than the explicit feedback-loop anchor of 3.	2 / 3
Progressive Disclosure	The skill is under 50 lines, single-purpose, and has no bundle files or external references to split out; the numbered list plus guardrail is well-organized, so per the simple-skills note progressive disclosure scores 3 without external file references.	3 / 3
	Total	10 / 12 Passed

Description

50%

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 third-person and names a concrete domain, but it omits any 'when to use' trigger guidance and relies on technical phrasing, leaving every dimension at the mid anchor.

Suggestions

Add an explicit 'Use when...' trigger clause, e.g. 'Use when improving Fallow's analysis results, validating tool output against source, or benchmarking Fallow against competitors.'

Broaden trigger terms with natural phrasings a user would actually say (e.g. 'check Fallow results', 'benchmark Fallow', 'find Fallow false positives') instead of only 'analysis accuracy' and 'source truth'.

List a couple of concrete actions to lift specificity toward 3, e.g. 'compare Fallow output against competitor tools, verify each disagreement against source, and keep only corrections that improve corpus-wide accuracy.'

Dimension	Reasoning	Score
Specificity	It names the domain ('Fallow analysis accuracy') and the means ('comparing it with competing tools and verified source truth across a stable real-world corpus'), but describes one composite action rather than listing multiple concrete actions, so it sits at the 'some actions, not comprehensive' anchor and not the multi-action level 3.	2 / 3
Completeness	It states what the skill does but gives no 'Use when...' clause or equivalent trigger guidance for when to invoke it, which per the guidelines caps completeness at 2 ('has what, but when is missing or only implied') and rules out 3.	2 / 3
Trigger Term Quality	Relevant domain keywords exist ('Fallow', 'analysis accuracy', 'competing tools', 'source truth'), but phrasing is technical/jargon-leaning and lacks natural user variations a person would actually say, so it matches 'some relevant keywords but missing common variations' rather than the broad natural-term coverage of 3 or the no-keyword case of 1.	2 / 3
Distinctiveness Conflict Risk	Naming the specific tool 'Fallow' gives it a niche, but the generic 'improve accuracy by comparing' framing and absence of explicit triggers could overlap with general testing/validation skills, placing it at 'somewhat specific but could still overlap' rather than the clearly-distinct level 3.	2 / 3
	Total	8 / 12 Passed

Validation

100%

Warnings & errors only

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: fallow-rs/fallow
Path: .agents/skills/conformance-loop/SKILL.md
Commit: 55944e8

Reviewed: about 18 hours ago

Table of Contents

Discovery Implementation Validation

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.