Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx skills add bunhine0452/Ocul-PM --skill redteamgit clone --depth 1 https://github.com/bunhine0452/Ocul-PMWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/bunhine0452/ocul-pm/redteam)<a href="https://agentmods.dev/skills/bunhine0452/ocul-pm/redteam"><img src="https://agentmods.dev/badge/skills/bunhine0452/ocul-pm/redteam/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/bunhine0452/ocul-pm/redteam"><img src="https://agentmods.dev/badge/skills/bunhine0452/ocul-pm/redteam.svg" alt="Reviewed on agentmods" width="80" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00058 | $0.00862 |
| Opus 5.5 | $0.00023 | $0.00345 |
| Sonnet 5.5 | $0.00012 | $0.00172 |
| Haiku 4.5 | $0.00006 | $0.00086 |
Grade A, and why
redteam scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured today.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
What it actually says
/oculpm:redteam — 플랜·일지의 반대편
"숏셀러처럼 써라"가 아니라 숏셀러가 하는 일을 하세요. 숏셀러는 글을 쓰지 않습니다 — 숫자를 다시 세고, 주장을 원재료와 대조하고, 틀리면 돈을 잃습니다. 이 스킬의 산출물은 논증이 아니라 재현된 반례입니다.
대상
인자가 없으면 활성 플랜 중 가장 최근에 갱신된 것. /oculpm:redteam <plan_id> 또는 일지 경로를 주면 그것.
절차 (형식 강제)
- 주장 목록 — 플랜의
**완료 조건:**과 완료([x]) 항목, 일지의 「검증」 칸에서 검증 가능한 주장만 뽑는다 (한 줄 = 한 주장). 의견·계획은 뽑지 않는다. - 원재료로 가기 — 주장마다 그것을 뒷받침한다는 원재료(테스트 파일·명령·로그·DB 행·git diff)를 찾는다. 요약·일지 문장·커밋 메시지는 원재료가 아니다.
- 직접 돌린다 — 테스트·린트·빌드·스크립트를 실제로 실행하고 exit code 와 숫자를 받는다. 돌릴 수 없는 주장은 「검증 불가」로 분류한다 — 그것도 발견이다.
- 가정을 뒤집는다 — "이 주장이 틀렸다면 어디서 먼저 드러나는가"를 하나 골라 그 자리를 본다 (경계값·빈 입력·동시 실행·다른 OS·오래된 데이터). 주장에 닿는 분야 둘~셋의 방법을 겹친다 — 통계라면 표본·분모를 다시 세고, 보안이라면 경계 입력을 넣고, 운영이라면 다른 환경에서 돌린다.
- 바깥 심판만 인용 — 네 판단("문제없어 보인다")은 결과에 쓰지 않는다. 결과 줄에는 명령·파일·숫자만.
결과 형식 (강제 — 자유 산문 금지)
REDTEAM <대상> · 주장 N · 반례 N · 검증 불가 N
반례
<file>:L<line> 또는 <명령>: <주장> ↔ <실제로 나온 것>. <재현: 한 줄 명령>.
검증 불가
<주장>: <왜 돌릴 수 없는가> → <무엇이 있으면 돌릴 수 있는가>.
유지된 주장
<주장>: <무엇이 그것을 보여줬나 (명령·숫자)>.
- 반례가 0 이면 "반례 0" 이라고 쓰고 유지된 주장에 근거를 전부 적는다 — 빈 보고는 금지.
- 반례는 고치지 않는다. 고칠지는 사용자가 정한다. 사용자가 고치라고 하면 그때
/oculpm:next로. - 결과로 별도 문서를 만들지 않는다 — 사용자가 요청할 때만
journal_write(type: chore, 제목 「레드팀 — 대상」)로 기록한다.
금지
- 역할극 문체("숏셀러로서 저는…"). 하는 일만 한다.
- 돌리지 않은 명령의 결과를 추정하기.
- 대상 플랜·일지의 수정. 읽기와 실행만.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- today Changed 80de640c0772
- yesterday First seen · 46 lines · 58 tokens per session scan A 703a84874432
redteam is a skill published in the GitHub repository bunhine0452/Ocul-PM (7 stars, last pushed today), licensed MIT. It adds 58 tokens to every session and 862 once invoked, about $0.0002 per session on Opus 5.5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-10-28.
Other skills, from other repositories
repair-prevention-assessment
A completion gate for fixes and self-repair work. It creates a machine-checkable assessment that separates evidence a current issue is closed from evidence that the fix will remain effective.
orchestrate-test-fix
Orchestrate phases 3 and 4 - find where the built agent is wrong or fragile (edge cases, overfitting, determinism, two-strategy comparison, one-input trace), then fix wrong rows one change at a time with a diff of the whole output. Use once the pipeline runs on the sample, and whenever output rows are wrong or weakly…
error-collection
A structured record of confirmed engineering mistakes, their root causes, and checks that prevent them from returning. It uses evidence from real application behavior, not just successful builds.
flaky-test-triage
Use when a test passes and fails intermittently, behaves differently by order, machine, runtime, timing, or retry, or when the user asks to reduce test flakiness. Gather repeatable evidence, isolate nondeterminism, and repair the causal dependency without weakening test intent.
test-debugger
Use when a test fails, flakes, hangs, or gives an unexpected result. Reproduce the failure with the narrowest package-local command, distinguish test defects from product defects and environment issues, and only implement a fix when the user authorizes changes.
test-failure-explorer
Import and query an existing Vitest 3.2 JSON report to explore test failures, file-level errors and original diagnostic evidence.