execute-probe

SkillDev tools

Lets your agent run a single security test or attack and record its outcome, evidence, and severity.

Available today. Use it from your connected AI after setup.

Add ahel to your AI once: Claude, ChatGPT, Cursor, Claude Code or Codex. Then ask it to use this.

Then ask your AI: use the execute-probe skill

About this skill

Execute one test/attack and record outcome, evidence, and severity. Tool use is runtime-selected.

What this skill tells your AI

The instructions your AI receives, as published by yogsoth-ai/de-anthropocentric-research-engine in skills/execute-probe/SKILL.md and read by ahel’s review.

Purpose

Execute one test or attack and record outcome, evidence, and severity.

Input contract

required: [probe_spec, target, success_and_failure_conditions]
optional: [tool_context, severity_scale]
constraints: [probe scope and interpretation rule must be fixed before execution]

Procedure

  1. Verify target and probe conditions.
  2. Run the probe and capture observations.
  3. Classify outcome against the predeclared conditions.

If probe observations and protocol deviations are recorded, consider classify-falsification-verdict as the next tactic.

Output contract

produces: [probe_record, observations, outcome, severity]
delta_fields: [findings, evidence_updates, uncertainties, decisions]

Quality gates

  • Preserve raw observation, execution context, and interpretation separately.

Failure and counterexamples

An unavailable tool or inconclusive observation is uncertainty, not a passing probe.

Provenance map

  • resolved: probe-execution

Signals

GitHub stars
501
Forks
41
Last commit
Sep 2026
Advanced
Catalog kind
skill
Key
execute-probe
Source
github.com/yogsoth-ai/de-anthropocentric-research-engine