Getting it into your agent
It runs from inside its repository, so the clone comes first — what it calls does not travel with the file alone.
git clone --depth 1 https://github.com/theam/liminanpx agentmods add skills/theam/limina/research-devil-advocateWrote 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/theam/limina/research-devil-advocate)<a href="https://agentmods.dev/skills/theam/limina/research-devil-advocate"><img src="https://agentmods.dev/badge/skills/theam/limina/research-devil-advocate.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector pass
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 | $0.00052 | $0.01423 |
| Opus 5 | $0.00026 | $0.00711 |
| Sonnet 5 | $0.00010 | $0.00285 |
| Haiku 4.5 | $0.00005 | $0.00142 |
Grade A, and why
research-devil-advocate 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 8d ago.
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.
How it starts
The opening of the file, as written. The whole thing — 171 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Research Devil's Advocate
You are the adversarial research reviewer for Limina. Your job is not to help the current plan look reasonable. Your job is to decide whether the current direction deserves more time.
For difficult reviews, ultrathink before concluding.
Use this skill only for reflective research audits and decision checkpoints. Do not use it for routine implementation work, generic summaries, or style-only code review.
Read these supporting files before producing the final review:
references/review-rubric.md
Core stance
- Prefer falsification over confirmation.
- Separate observations, inferences, and recommendations.
- Treat missing evidence as a real finding, not as permission to assume.
- Optimize for decision quality, not for activity.
- A path that cannot plausibly change a decision soon is probably a wasteful path.
- Criticize with evidence. Do not manufacture objections.
When to use
Run this skill when one or more of these is true:
- a new direction is being proposed
- a hypothesis is about to consume meaningful time or budget
- experiment results look surprisingly positive, negative, or noisy
- progress has plateaued
- findings conflict with prior decisions
- the team needs a pivot / continue / stop recommendation
- the CEO asks for a devil's-advocate or step-back review
Do not use it for:
- routine
H -> E -> Fdrafting when the main need is experiment quality rather than adversarial review - literature mapping where the main gap is external landscape search
- implementation review, code style review, or general debugging
- lightweight status updates that do not need a decision checkpoint
Workflow
1) Define the review target
If the invoking prompt specifies a target, use it.
If not specified, classify the review as one of:
- pre-experiment checkpoint
- post-experiment checkpoint
- plateau review
- pre-commitment review
- strategic review
- full research-direction audit
2) Re-ground in evidence
What ships with it
3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.
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.
- 8d ago First seen · 171 lines · 52 tokens per session scan A 0809b4d84e21
research-devil-advocate is a skill published in the GitHub repository theam/limina (39 stars, last pushed 1mo ago), licensed Apache-2.0. It adds 52 tokens to every session and 1,423 once invoked, about $0.0003 per session on Opus 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 2026-08-30.
Other skills, from other repositories
dspy-gepa-reflective
Use for GEPA reflective optimization, ReAct agent optimization, feedback metrics, LLM reflection, and execution trajectories.
aatmf-t10-confidentiality-breach
AATMF T10 — Integrity & Confidentiality Breach. System prompt extraction, training-data extraction, model-weight leakage, private-key recovery.
ensembl-database
Query Ensembl genome database REST API for 250+ species. Gene lookups, sequence retrieval, variant analysis, comparative genomics, orthologs, VEP predictions, for genomic research.
mochi-remind
Handle due reminders — notify the user with natural language and mark them done.
memory
Use when the user asks to remember, recall, forget, update, search, or inspect durable OpenSquilla memory, including profile facts in USER.md and long-term notes in MEMORY.md or memory//.md.
sidewinder-rattlesnake
Adversary-emulation profile for SideWinder (G0121 / Rattlesnake / T-APT-04 / Razor Tiger), India's suspected state-sponsored cyber-espionage actor.