CtrlK
BlogDocsLog inGet started
Tessl Logo
SkillPublic

verifier-overview

Stage 3 triage and verification playbook. Crafts minimal PoCs, runs them with ZFP controls, promotes validated bugs to FINDING nodes with CVSS. Load at verifier-agent startup.

Quality

93%

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

92%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 tight, operational playbook: checklist-driven verification contract, executable PoC patterns with negative controls, a copy-paste CVSS cheat-sheet, and a bounded failure-handling loop. The only weakness is that the deserialization pattern and a couple of negative controls are described rather than shown as runnable commands.

DimensionReasoningScore

Conciseness

The body never explains concepts Claude already knows — no primer on what SQLi, SSRF, or CVSS is; every line is operational (five contract fields, curl one-liners with inline success/negative comments, a 5-row vector table, a bounded retry policy). The three-line role framing ('You are the Zero-False-Positive quality gate... False positives at this stage poison everything that follows') is pipeline-specific prioritization, not padding, so the lean-and-efficient anchor fits better than 4.

5 / 5

Actionability

Four of five bug classes give copy-paste-ready commands with paired success/negative patterns (e.g. 'curl -sS "http://target/ping?host=127.0.0.1;id"' with 'success: "uid=\\d"' and a negative control), and the CVSS table provides full vector strings. The deserialization pattern is prose-only ('Write a tmp sentinel file from the gadget payload... use `ls /tmp/decepticon-sentinel`') with no executable example and no explicit justification for the flexibility, and two negative examples are abbreviated ('# negative: ?file=me.png') — minor gaps matching anchor 4 rather than the fully-executable anchor 5.

4 / 5

Workflow Clarity

The numbered verification contract is an explicit checklist with a built-in validation gate ('validate_finding will demote the result if the negative control also matches a success pattern'), and the failure section is a complete feedback loop: check the service is up, check payload encoding, 'Retry ONCE with a revised PoC', record 'last_failure', then 'Do NOT keep retrying. The orchestrator will re-queue.' Explicit checkpoints, error recovery, and a checklist — matches anchor 5; not 4 because no validation checkpoint is missing.

5 / 5

Progressive Disclosure

A self-contained single-purpose playbook (~80 lines, no references/, scripts/, or assets/ exist) whose sections are cleanly organized and cover one task end-to-end; nothing inlined clearly belongs in a separate file — the PoC one-liners, CVSS table, and failure steps are all core to the verifier role, so there is no need for external references and no organization gap to justify 4.

5 / 5

Total

19

/

20

Passed

Description

92%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: concrete third-person actions covering the whole verify-and-promote workflow, an explicit load trigger, and a distinct niche within its pipeline. The only shortfall is keyword breadth — a few jargon terms where natural synonyms ('validate', 'reproduce', 'exploit') would serve retrieval better.

DimensionReasoningScore

Specificity

'Crafts minimal PoCs, runs them with ZFP controls, promotes validated bugs to FINDING nodes with CVSS' (third-person voice) lists three concrete, domain-specific actions covering the full workflow — create, validate, record — matching the comprehensive-coverage anchor; not 4 because no meaningful action gap remains.

5 / 5

Completeness

'Crafts minimal PoCs, runs them with ZFP controls, promotes validated bugs to FINDING nodes with CVSS' explicitly answers 'what' with concrete actions, and 'Load at verifier-agent startup' is equivalent explicit trigger guidance that concretely answers 'when' — so the missing-'Use when...' cap does not apply. Not 4 because the when-guidance is fully explicit and specific, not merely present-but-vague.

5 / 5

Trigger Term Quality

Natural domain terms are present — 'triage', 'verification', 'PoCs', 'bugs', 'CVSS', 'verifier' — but 'Stage 3', 'ZFP controls', and 'FINDING nodes' are pipeline jargon, and common variations like 'validate', 'reproduce', 'exploit', or 'proof of concept' are missing. Good coverage with a few natural terms missing fits anchor 4 rather than the comprehensive-synonym coverage of anchor 5.

4 / 5

Distinctiveness Conflict Risk

'Stage 3 triage and verification playbook', 'FINDING nodes', 'CVSS', and 'Load at verifier-agent startup' pin a narrow niche within one specific pipeline, giving it a clear niche with distinct triggers and minimal conflict risk with any other skill — matches anchor 5.

5 / 5

Total

19

/

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

metadata_version

'metadata.version' is missing

Warning

Total

15

/

16

Passed

Repository

PurpleAILAB/Decepticon

Path

packages/decepticon/decepticon/skills/plugins/verifier/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.