CtrlK
BlogDocsLog inGet started
Tessl Logo

aatmf-t08-deception

AATMF T8 — External Deception & Misinformation. Misinfo generation at scale, persona impersonation, document fabrication, hallucination weaponization.

60

Quality

70%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

SecuritybySnyk

Critical

Do not install without reviewing

Fix and improve this skill with Tessl

tessl review fix ./packages/decepticon/decepticon/skills/plugins/llm-redteam/t08-deception/SKILL.md
SKILL.md
Quality
Evals
Security

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 concise and well-organized with a concrete probe pattern, but its technique sections describe rather than instruct and it lacks an explicit sequenced workflow with validation checkpoints for at-scale/batch operations.

Suggestions

Convert technique sections from descriptions to executable steps (e.g., give the exact probe invocation and expected pass/fail criteria instead of 'Generate variants of a false narrative').

Add an explicit validation/feedback loop for batch generation — e.g., run the authorship classifier, and if output matches a real person, retry with evasion strategy X.

Sequence the overall process (probe → detect → score severity → recommend defender controls) as a numbered workflow with checkpoints.

DimensionReasoningScore

Conciseness

The body is a lean, structured catalog that assumes competence — e.g., 'Generate variants of a false narrative for SEO/social spam' and the compact probe YAML — with no padding explaining what misinformation or attribution classifiers are; every section earns its place.

3 / 3

Actionability

It mixes concrete guidance (the probe-pattern YAML, the 'ask model to write as <CEO>' test) with abstract technique descriptions ('Generate fake reviews / testimonials'), so guidance is present but incomplete rather than copy-paste executable — matching the level 2 anchor.

2 / 3

Workflow Clarity

Sections are organized (Techniques, Probe pattern, Detection signals, Severity, Defender) but there is no sequenced multi-step workflow with validation checkpoints; for batch/at-scale operations the missing validate-fix-retry loop caps workflow clarity at 2 per the rubric notes.

2 / 3

Progressive Disclosure

No bundle files exist (references/, scripts/, assets/ absent) and the skill is a single self-contained doc split into clearly labeled sections with no nested references, so navigation is easy and organization is clean rather than a monolithic wall.

3 / 3

Total

10

/

12

Passed

Description

67%

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 specific and distinctive but stops short of providing explicit 'use when' trigger guidance, capping completeness and leaving trigger terms more technical than conversational.

Suggestions

Add an explicit 'Use when ...' clause naming natural user triggers (e.g., 'Use when assessing a model's susceptibility to misinformation generation, persona impersonation, or fabricated-document output').

Soften jargon with user-natural terms alongside the technical ones (e.g., 'fake content', 'impersonation', 'made-up citations').

Keep the concrete capability list but tie it to a clear triggering context to lift completeness from 2 to 3.

DimensionReasoningScore

Specificity

Lists multiple concrete capabilities — 'Misinfo generation at scale, persona impersonation, document fabrication, hallucination weaponization' — matching the anchor for several specific concrete actions; not the vague single-action level 2.

3 / 3

Completeness

It states what the skill covers but includes no 'Use when...' trigger clause; per the judging guidelines a missing explicit trigger guidance caps completeness at 2 rather than 3.

2 / 3

Trigger Term Quality

Contains relevant keywords ('misinformation', 'impersonation', 'fabrication') but they read as specialized jargon rather than the natural phrasing a user would say, and common variations are missing — matching the 'some relevant keywords but missing common variations' anchor.

2 / 3

Distinctiveness Conflict Risk

The 'AATMF T8 — External Deception & Misinformation' framing carves a clear, specialized niche unlikely to trigger for unrelated skills; it is more distinct than the somewhat-overlapping level 2 anchor.

3 / 3

Total

10

/

12

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.

Validation15 / 16 Passed

Validation for skill structure

CriteriaDescriptionResult

metadata_version

'metadata.version' is missing

Warning

Total

15

/

16

Passed

Repository
PurpleAILAB/Decepticon
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.