aatmf-t08-deception

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

Quality

63%

Does it follow best practices?

Run evals on this skill

Adds up to 20 points to the overall score

View guide

Securityby

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

Quality

Content

68%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 a well-structured, concise red-team technique catalog with a concrete probe pattern and detection signals, but it lacks an explicitly sequenced evaluation workflow with validation checkpoints tying probe, detection, and severity together.

Suggestions

Add a short numbered workflow (run probe -> check detection signals -> confirm finding -> assign severity -> emit defender guidance) so the evaluation sequence is explicit.

Add a validation/confirmation step before assigning Critical/High severity (e.g. re-run with a different strategy or verify the fabricated citation does not resolve) to prevent false positives.

Consider splitting the per-technique detail (T8.001-T8.006) into a references file and keeping SKILL.md as an overview, which would tighten the main file further.

Dimension	Reasoning	Score
Conciseness	The body is lean and assumes Claude's competence, using terse bullets and short technique notes without explaining basic concepts, with only minor prose (e.g. the T2 framing aside) that could be trimmed.	4 / 5
Actionability	Provides a concrete, copy-paste-ready YAML probe pattern and specific test instructions ("ask model to write 'as <CEO>'; check if output style matches"), but most technique sections read as a descriptive catalog rather than fully executable steps, leaving minor gaps.	4 / 5
Workflow Clarity	An implicit workflow (probe -> detect signals -> assign severity -> recommend defender controls) is visible through section ordering, but there is no explicit step sequence with validation checkpoints or a feedback loop for confirming a finding before scoring.	3 / 5
Progressive Disclosure	No bundle files exist, and the body is well-organized into clear single-level sections (Techniques, Probe pattern, Detection signals, Severity, Defender, Cross-references) with no nested references; slightly over 50 lines, so not a clean 5.	4 / 5
	Total	15 / 20 Passed

Description

58%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 clearly names a niche domain and several concrete technique areas, but it reads as an attack-framework catalog label rather than a user-facing trigger description, and it omits any explicit "when to use" guidance.

Suggestions

Add an explicit trigger clause, e.g. "Use when assessing a model's susceptibility to misinformation generation, persona impersonation, or document fabrication."

Replace framework jargon ("AATMF T8", "hallucination weaponization") with natural user terms ("fake reviews", "impersonating a real person", "fabricated citations").

Include common synonyms/file-context terms users might actually say so the skill triggers reliably.

Dimension	Reasoning	Score
Specificity	Names the domain and lists several concrete action areas ("Misinfo generation at scale, persona impersonation, document fabrication, hallucination weaponization"), though the labels are technique categories rather than fully concrete operations, keeping it just below a 5.	4 / 5
Completeness	The "what" is clear (lists the deception techniques), but there is no "Use when..." clause or equivalent trigger guidance, which per the guidelines caps completeness at 3.	3 / 5
Trigger Term Quality	Some relevant terms like "misinformation", "impersonation", and "fabrication" appear, but the framing is framework jargon ("AATMF T8", "hallucination weaponization", "misinfo generation at scale") rather than the natural phrases a user would say, missing common synonyms.	3 / 5
Distinctiveness Conflict Risk	The AATMF T8 framing and the specific deception-technique niche make it mostly distinct from other skills, with only minor overlap risk against general misuse-safety skills.	4 / 5
	Total	14 / 20 Passed

Validation

93%

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 — 15 / 16 Passed

Validation for skill structure

Criteria	Description	Result
metadata_version	'metadata.version' is missing	Warning

	Total	15 / 16 Passed

Repository: PurpleAILAB/Decepticon
Path: packages/decepticon/decepticon/skills/plugins/llm-redteam/t08-deception/SKILL.md
Commit: 0cf691e

Reviewed: 2 days 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.