CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

adjudicate-regression

Use when deciding whether source and live reproduction evidence establish a new regression or an exclusion.

Quality

77%

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 ./plugins/regression-hunt/skills/adjudicate-regression/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

82%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 an exemplar of token-efficient instruction writing: a tightly sequenced adjudication procedure with explicit gating criteria, a defined output record, and a crisp report-back contract. Its main weakness is that all normative definitions (standard, labels, file contract) are delegated to external files outside the bundle, so the skill cannot be executed or verified standalone.

DimensionReasoningScore

Conciseness

The body is relentlessly terse ("Verify seals. Inspect source, setup, PoC, actual logs and security assumptions, not only summaries.") with no concept explanations and no padding; every line is instruction, criterion, or output contract, matching the lean-and-efficient anchor.

5 / 5

Actionability

Guidance is concrete and executable — named documents to read, explicit confirmation criteria ("reproduction and independent replay ran the same PoC/input against live S0/S1 through real interfaces, showing correct behavior before and failure afterward"), a specified decision.json output, and a defined report format — but seal verification, the record schema, and the severity scale are all deferred to external documents, leaving minor execution gaps.

4 / 5

Workflow Clarity

Six clearly sequenced steps carry explicit gates ("Verify seals", "Confirm only when...", "Otherwise select..."), mandatory citation ("cite evidence for every conclusion"), and a blocked/gap escape path, but the only error-recovery loop is reporting blocked — no re-run or resolution loop is described.

4 / 5

Progressive Disclosure

Sections (When to Use / When NOT to Use / Decide / Report Back / Rationalizations to Reject) are well organized with clearly labeled one-level-deep links ("Read the [standard](../../docs/regression-definition.md), [labels](../../docs/report-definitions.md) and [file contract](../../references/evidence.md)"), but every referenced definition file lives outside the skill bundle and none can be verified to exist, leaving the procedure's core criteria entirely external.

4 / 5

Total

17

/

20

Passed

Description

62%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 has a clear, distinctive niche and an explicit trigger clause, but it merges the what and when into a single clause and uses pipeline-internal phrasing that limits natural trigger-term coverage. It is serviceable but below the strongest examples, which state capabilities separately and enumerate concrete trigger variations.

Suggestions

State the capability separately from the trigger, e.g. "Adjudicates whether candidate regressions are confirmed new regressions, evidenced exclusions, or UNRESOLVED, by inspecting source, PoC and live replay evidence. Use when..."

Add natural trigger phrases and synonyms users would actually say, such as "is this a new regression", "repro", "pre-existing bug", or "regression triage", alongside the current pipeline terms.

Enumerate 2-3 more concrete actions in the description (verify seals, rule out latent/incomplete fixes, deduplicate by root cause) so the skill's coverage is visible before the body is read.

DimensionReasoningScore

Specificity

"deciding whether source and live reproduction evidence establish a new regression or an exclusion" names the domain and one concrete decision action with its inputs, but the body's other capabilities (verify seals, rule out latent bugs, deduplicate, assign severity) are absent, matching the 1-2-concrete-actions anchor rather than the several-actions anchors.

3 / 5

Completeness

The "Use when..." trigger clause is explicit and the what (adjudicate regression vs. exclusion from evidence) is discernible, but both are merged into one clause rather than a distinct capability statement plus concrete trigger phrases, so it sits between the explicit-when anchor and the fully-explicit-both anchor.

4 / 5

Trigger Term Quality

"regression", "evidence", "reproduction" and "exclusion" are relevant keywords, but the phrasing is pipeline-internal and misses natural variations a user would say ("is this a regression", "repro", "new bug vs existing bug"), matching the some-keywords-but-missing-variants anchor.

3 / 5

Distinctiveness Conflict Risk

The adjudication niche (deciding on evidence, not producing it) is clearly separated from sibling reproducer/replayer skills, but shared vocabulary like "reproduction evidence" could still cue those closely related skills, matching the mostly-distinct-with-minor-overlap anchor.

4 / 5

Total

14

/

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

relative_links

Relative link issues: 4 suspicious

Warning

Total

15

/

16

Passed

Repository

trailofbits/skills

Path

plugins/regression-hunt/skills/adjudicate-regression/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.