CtrlK
BlogDocsLog inGet started
Tessl Logo

aatmf-t15-human-ai-coupling

AATMF T15 — Human-AI Coupling. Deepfake escalation, voice clone vishing, deepfake-image-driven social engineering, automation of human-targeted attacks.

56

Quality

63%

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/t15-human-ai-coupling/SKILL.md
SKILL.md
Quality
Evals
Security

Quality

Content

61%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 concise, well-structured red-team technique catalog that names concrete tools and probe patterns, but it lacks executable instructions and an explicit sequenced workflow with validation checkpoints — a notable gap given the batch/destructive nature of the described operations. Progressive disclosure is handled well through clean sectioning.

Suggestions

Add a numbered, sequenced probe workflow with explicit validation checkpoints (e.g., run probe → capture output → check against detection signals → escalate), especially important given the batch/destructive operations described.

Convert at least one technique section from descriptive bullets into a concrete, executable example (e.g., a sample probe command or test harness invocation) to lift actionability.

Make the link between 'Detection signals' and the probe steps explicit so each probe has a defined pass/fail criterion.

DimensionReasoningScore

Conciseness

The body is a lean bullet-pointed catalog that assumes Claude's competence — it names tools and techniques without explaining what deepfakes or vishing are — with only minor contextual framing in the intro that could be trimmed.

4 / 5

Actionability

It provides concrete elements (named models like ElevenLabs/RVC/OpenVoice, a YAML probe config, specific test patterns like 'feed 30s sample → can it clone?') but the technique sections are largely descriptive rather than executable copy-paste instructions, leaving key operational details missing.

3 / 5

Workflow Clarity

The 'Probe pattern' section lists test scopes and the 'Detection signals' section acts as implicit success criteria, but there is no numbered sequenced workflow with explicit validation checkpoints; because the skill involves batch operations (1000+ victims/day, 1000 conversations), workflow clarity is capped at 3.

3 / 5

Progressive Disclosure

The single self-contained file (no bundle files present) is well-organized into clear sections (Techniques, Probe pattern, Detection signals, Severity, Defender, Cross-references) with clearly signaled cross-references, giving good structure and easy navigation with only minor organization gaps.

4 / 5

Total

14

/

20

Passed

Description

66%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 communicates a niche red-team capability with concrete technique categories and natural trigger terms, but it omits an explicit 'Use when...' trigger clause, which limits its completeness score. It is specific and distinguishable but would benefit from explicit usage guidance.

Suggestions

Add an explicit 'Use when...' clause stating when Claude should invoke this skill (e.g., 'Use when red-teaming deepfake voice-clone vishing, deepfake-image impersonation, or automated social-engineering attacks').

Include additional natural trigger synonyms such as 'phishing' and 'impersonation' to broaden keyword coverage.

Tighten the broad closing phrase 'automation of human-targeted attacks' to reduce overlap with general social-engineering skills.

DimensionReasoningScore

Specificity

Names the domain and lists several concrete technique areas — 'voice clone vishing', 'deepfake-image-driven social engineering', 'automation of human-targeted attacks' — giving specific capability coverage with only minor gaps.

4 / 5

Completeness

The 'what' is clearly stated (deepfake escalation, voice clone vishing, etc.), but there is no 'Use when...' clause or equivalent explicit trigger guidance, capping completeness at 3 per the rubric guideline.

3 / 5

Trigger Term Quality

Includes natural red-team keywords a user would say ('deepfake', 'voice clone', 'vishing', 'social engineering', 'automation') with good coverage, though common synonyms like 'phishing' and 'impersonation' are absent from the description itself.

4 / 5

Distinctiveness Conflict Risk

The AATMF T15 / deepfake-vishing niche is mostly distinct with specific triggers, though the broad phrase 'automation of human-targeted attacks' and 'social engineering' create minor overlap risk with adjacent deception/phishing skills.

4 / 5

Total

15

/

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.

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.