CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

workflow-guardrail-candidates

Binary calibration pack for recurring workflow guardrail recommendations.

Quality

48%

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 ./codex/agents/evals/workflow-guardrail-candidates/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

37%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 lean and cleanly organized as a directory map, but it provides zero actionable guidance — no instructions, commands, or workflow for using the eval pack — and most of its referenced paths are dead while real bundle files go unmentioned. A reader learns where things would be, not what to do.

Suggestions

Add a short 'How to use' section with concrete steps/commands: how to run the eval runner, and how to route a guardrail claim to a case in evals.yaml and judge it against its acceptance criteria.

Align the Layout section with the actual bundle: `evals/`, `tessl.json`, and `workflow-guardrail-candidates.json` do not exist; instead document `references/contract.yaml`, `references/task-profile.json`, and `references/scorer-calibration/`.

Include a validation checkpoint for interpreting results, e.g. what to do when a candidate's acceptance criteria fail or when only the skill name (no route proof) is available.

DimensionReasoningScore

Conciseness

The ~13-line body is a lean layout map with zero padding and no explanations of concepts Claude already knows — "every token earns its place". Not below 5 because there is no over-explanation or filler anywhere.

5 / 5

Actionability

The body "only describes rather than instructs": it states that "`references/evals.yaml` holds the claim-to-case routing and calibration expectations" but gives no steps, commands, or code for actually producing or evaluating a recommendation. Not 2 because there is not even a high-level hint about how to use the pack.

1 / 5

Workflow Clarity

No sequence or steps of any kind appear — the body is purely a file inventory ("`evals/` contains per-case task and criteria files"). This matches 'Steps missing or incoherent; no sequence; no validation for risky operations'; not 2 because there is no rough sequence to evaluate.

1 / 5

Progressive Disclosure

The Layout section is a clean, well-signaled one-level overview and `references/evals.yaml` exists and matches its description, but three of the four referenced paths (`evals/`, `tessl.json`, `workflow-guardrail-candidates.json`) do not exist in the bundle, while existing files (`references/contract.yaml`, `references/task-profile.json`, `references/scorer-calibration/`) go undocumented. This exceeds 'minor organization gaps' so not 4, but the SKILL.md itself is well-organized so it sits above 2.

3 / 5

Total

10

/

20

Passed

Description

32%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 names a niche domain but lists no concrete actions, includes no trigger guidance ('Use when...'), and leans on internal jargon ('binary calibration pack') rather than natural user phrasing. It is unlikely to be selected when needed because nothing tells Claude or the user when to reach for it.

Suggestions

Add an explicit trigger clause, e.g. 'Use when evaluating a recurring workflow guardrail recommendation or deciding whether a guardrail candidate should be implemented.'

State 1-2 concrete actions instead of an artifact label, e.g. 'Routes workflow guardrail candidates to calibration cases in references/evals.yaml and scores each as implement/reject with evidence.'

Replace internal jargon like 'Binary calibration pack' with natural phrasing users would actually say, such as 'workflow guardrail recommendations' or 'guardrail candidate review'.

DimensionReasoningScore

Specificity

"Binary calibration pack" and "recurring workflow guardrail recommendations" name the domain, but the description contains no concrete actions — no verbs describing what the skill actually does — matching the anchor 'Names the domain but actions are minimal or generic'. Not 3 because no 1-2 concrete actions are listed at all.

2 / 5

Completeness

The 'what' is a vague artifact label ("Binary calibration pack") rather than a clear statement of capability, and the 'when' is entirely absent — no 'Use when...' clause. This fits 'Has a vague what and no when'; not 3 because the 'what' is not clear and concrete, and per guidelines a missing 'Use when' clause caps completeness at 3, which this description falls below.

2 / 5

Trigger Term Quality

"workflow guardrail" is a plausible user term, but "Binary calibration pack" is internal jargon no user would naturally say, and there are no synonyms or natural trigger phrases. Fits 'One or two generic keywords; missing the natural phrases users say'; not 3 because the only relevant keyword set is jargon-heavy.

2 / 5

Distinctiveness Conflict Risk

"recurring workflow guardrail recommendations" carves a clear niche that is unlikely to collide with typical skills — mostly distinct with minor overlap risk with other eval/calibration pack skills. Not 5 because no distinct trigger terms are stated, and not 3 because the domain is narrower than 'somewhat specific'.

4 / 5

Total

10

/

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

jscraik/Agent-Skills

Path

codex/agents/evals/workflow-guardrail-candidates/SKILL.md

Revision

2920192

Reviewed

Table of Contents

DiscoveryImplementationValidation

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.

Claim skill