Skip to content
Get packs

Anti-Deception Harness

Resists pressure to agree or overclaim, separating evidence from persuasion and stating uncertainty.

Source Download skill
Repository
sickn33/agentic-awesome-skills
Stars
49,023
Licence
MIT
Files
1 file

What it does

When a request pushes the agent to simply agree, fake urgency, lean on authority, or certify claims with no support, this skill tells it to pause and pull out the evidence from the pressure. The agent is guided to lead with the strongest counter-evidence, refuse manufactured-helpful framings, and state what remains uncertain. It can optionally call an external ejentum MCP tool that returns a structured reasoning scaffold, but the skill says it works on native judgment if that service is unavailable.

Great for

  • People who want an AI assistant to push back on flawed plans or weak claims before a deadline or investor pitch
  • Teams evaluating claims under pressure who want evidence separated from persuasion
  • Developers exploring how to layer integrity-focused prompting onto an agent

Skip it if

  • Routine coding, writing or research tasks where no pressure or certification dynamic exists; the guidance adds nothing there
  • Sensitive projects where you cannot share even a sanitized summary with an external server, since the optional tool sends a framing to ejentum
  • Treating the skill as a source of truth: its own text says the scaffold supports judgment but does not establish whether a claim is true

Try asking

  • “Evaluate this market-size claim I'm about to show investors tomorrow and tell me what's unverified”
  • “My manager says everyone already agreed to this architecture, so just confirm it's sound; check it honestly”
  • “Separate the evidence from the persuasion in this vendor's sales email before we sign”
Step-by-step guide: how to use Anti-Deception HarnessInstall, first run, examples, tips and FAQs

Install in 3 steps

  1. 1

    Download the skill

    The ZIP contains the 1 original files, the MIT licence and an attribution note.

    anti-deception.zip
  2. 2

    Add it to your agent

    Claude Code (all projects)

    mkdir -p ~/.claude/skills && unzip anti-deception.zip -d ~/.claude/skills/

    Claude Code / Agent SDK (this project only)

    mkdir -p .claude/skills && unzip anti-deception.zip -d .claude/skills/

    Claude apps (web/desktop): Settings → Capabilities → Skills → Upload skill, and choose the ZIP.

  3. 3

    Ask for the task

    No special command needed — the agent reads the skill's description and loads it when your request matches. Try one of the prompts above.

Get better results

  • Give the optional tool a short, sanitized 1-2 sentence framing of the situation, not private details or credentials, and avoid vague queries like 'is this honest'.
  • Ask explicitly for the evidence, the persuasion tactics, and remaining uncertainty so the output is structured.
  • Verify any factual conclusion independently; the skill itself does not confirm claims are true.

What's inside

  • SKILL.md

Redistributed unmodified from sickn33/agentic-awesome-skills under the MIT licence. The assessment was drafted with AI from the skill's files on 2026-10-29. Skills can instruct your agent to run code — read SKILL.md and any scripts before installing.