interrogate
/interrogate
Use for \"interrogate\", \"adversarial review\", \"multi-model review\", \"challenge this\", \"stress test this code\", \"find blind spots\", or \"tear this apart\". You attack a change or design as an adversary, and only findings proven by a run survive.
- Author
- KyrosHendrix ↗
- Repository
- skills ↗
- Licence
- MIT
- Last repository change
- 03.11.2026
Assessment and security
- Files in this package
- 2
- Scripts
- 0
- HTML comments
- 0
- URLs in SKILL.md
- 0
No signals detected. Repository changed on 03.11.2026, without references to retired models or APIs.
MIT, open: use and modify while retaining author attribution.
What it does and when to use it
Use for \"interrogate\", \"adversarial review\", \"multi-model review\", \"challenge this\", \"stress test this code\", \"find blind spots\", or \"tear this apart\". You attack a change or design as an adversary, and only findings proven by a run survive.
Triggers from the description
Original author description
Use for \"interrogate\", \"adversarial review\", \"multi-model review\", \"challenge this\", \"stress test this code\", \"find blind spots\", or \"tear this apart\". You attack a change or design as an adversary, and only findings proven by a run survive.
How it works
- Frame
- Attack
- Verify
- Output
SKILL.md preview
--- name: interrogate description: "Use for \"interrogate\", \"adversarial review\", \"multi-model review\", \"challenge this\", \"stress test this code\", \"find blind spots\", or \"tear this apart\". You attack a change or design as an adversary, and only findings proven by a run survive." disable-model-invocation: true --- # Interrogate This skill runs inline, so you do every step yourself. Spawn subagents only when the user asks for it, per the Delegation section of [HARNESSES.md](../HARNESSES.md), which also maps tool and path terms. Act as an adversary and try to break a change or a design. A finding survives only when a run proves it. Report the findings. Do not apply fixes. ## Frame The scope is what the user names. Otherwise it is the branch diff from its merge base with the base branch, through the working tree. State the intent with the [in-your-own-words](../in-your-own-words/SKILL.md) skill, from the user's message, the commits, and the PR description. Ask the user when the intent is unclear. ## Attack Work from the scope and the intent, and follow this brief: > Break this change. Prove each finding with a run against the real code, such as a failing test, a repro script, or command output that shows the wrong behavior. For a design or plan, run a prototype or probe the current tree. Keep repros out of the scope's files, in a scratch location or your own worktree. Report each finding with a title, the exact command, the output it printed, and the output the intent requires. Leave out anything you did not run. That covers style, naming, structure, and risks that need a future change to matter. No findings is a valid result. Take at least two distinct angles of attack, one after another, such as bad inputs and edge cases first, then state, concurrency, and failure paths. Write down each angle's findings before starting the next. **Dispatch** (only when the user asks): run Phase A steps 3 and 4 and Phase B of the [arena](../arena/SKILL.md) skill instead, with three changes. Copy the scope's uncommitted changes into each worktree. Spawn each adversary as a subagent that can write and run code. Give each adversary the scope, the intent, and the brief above in place of the arena task and rationale. ## Verify Verify each finding per arena Phase F by rerunning its command. Keep a finding when the output shows the failure and the failure breaks the intent. Drop the rest. Merge findings that share a root cause. Skip arena Phases C, D, and E. Different angles find different failures, so the verdict keeps every verified finding, not one angle's list. ## Output - **Intent**: the restatement. - **Findings**: for each verified finding, the command, its output, how it breaks the intent, and the angle or adversary that found it. Say "no verified findings" when none survive. - **Dropped**: for each angle or adversary, the count of dropped findings and why.
Files in this package (2)
- SKILL.md 3 KB
- agents/openai.yaml 1 KB
Versions
| Repository | Score | Licence |
|---|---|---|
| This page · KyrosHendrix/skills | 90 | MIT |
| hootener/brain | 100 | MIT |
| requie/endostack | 100 | MIT |
| Sup-Film/pstack-omp-lean | 100 | MIT |
| RIJUDEVx/agent-skills | 90 | MIT |
| chmonitor/chmonitor | 90 | GPL-3.0 |
| richardochristjia/skills | 90 | MIT |
| chhoumann/t3-pstack | 90 | MIT |
| backnotprop/pstack | 90 | MIT |
Database snapshot: 13.11.2026 · package version 17d48465a79a (SHA-256 digest of the entire package)
Downloads in the last 30 days: 0 · ranking
Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.
Similar skills
- 95adversarial-review
Try to break a change or design and keep only findings proven by a run. Fresh adversary agents in their own worktrees attack the change against a restated intent; every finding is rerun before it survives. Also covers blast radius (what a change breaks elsewhere, with a proof… - 100comms-interrogate
Multi-reviewer critique of a diff, run as parallel read-only subagents from distinct angles, then synthesized into one verdict. Use for "interrogate", "adversarial review", "challenge this", "stress test this code", "find blind spots", or "tear this apart". - 100lie-detector
Toph's lie detector for multi-model adversarial review. She feels a heartbeat through the ground and knows when someone is lying, so a change that only claims to work does not fool her. Use for interrogate, /interrogate, adversarial review, multi-model review, "challenge this",… - 90mroowa-interrogate
Adversarially review code changes with one read-only reviewer and a parent judgment pass. Use for interrogate, challenge this code, stress test this diff, find blind spots, or a requested adversarial code review.