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/terrylica/cc-skillsnpx agentmods add skills/terrylica/cc-skills/m3Wrote 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/terrylica/cc-skills/m3)<a href="https://agentmods.dev/skills/terrylica/cc-skills/m3"><img src="https://agentmods.dev/badge/skills/terrylica/cc-skills/m3.svg" alt="Measured on agentmods" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium Prompt Injection · line 45 Large whitespace padding was detected (a block of blank lines or a long run of spaces). This can push injected instructions below or to the right of the visible area so a human reviewer never sees them while the agent still reads them. Manual review of the hidden content is recommended.Fix: Remove the large whitespace padding (blank-line blocks or long space runs) and review any content hidden below or to the right of it. Keep skill files compact and reviewable so no instructions can be
- medium Agent Snooping · line 96 Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
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 | $0.00162 | $0.01696 |
| Opus 5 | $0.00081 | $0.00848 |
| Sonnet 5 | $0.00032 | $0.00339 |
| Haiku 4.5 | $0.00016 | $0.00170 |
Grade A, and why
m3 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 3d 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 — 100 lines — stays where its author put it; the contents beside it link to each section on GitHub.
MiniMax-M3 — Production Wiring (empirical)
The M3 companion to ../minimax/SKILL.md (M2.7). Every claim here was
live-probed 2026-06-01 (fast subset re-verified 2026-06-23) on the Plus-High-Speed key. Full evidence + copy-paste snippets:
../../references/M3-EMPIRICAL.md.
Self-Evolving Skill: improves through use. If a flag stopped working, a limit moved, or the docs caught up with reality — fix this file +
references/M3-EMPIRICAL.mdimmediately, don't defer. Re-verify with the scripts below before changing a documented fact.
The one rule: default to reasoning_split: true
M3 still emits <think>…</think> inside content by default (same footgun as M2.7).
Setting reasoning_split: true moves the reasoning into a separate reasoning_content /
reasoning_details field and leaves content clean — no regex stripping. This is the
chosen default profile for everything migrating off M2.7-highspeed.
body = {
"model": "MiniMax-M3",
"messages": messages,
"max_tokens": 4096, # >= 1024 — thinking consumes budget before visible content
"temperature": 0.2,
"reasoning_split": True, # clean content; reasoning in reasoning_content/_details
}
answer = resp["choices"][0]["message"]["content"] # already clean — display directly
Need M2.7-highspeed-class speed on short/simple tasks? Add "reasoning": "disabled"
(≈2× fewer tokens, ≈2× faster) — and keep the M2.7 <think> strip as a safety net, since
"disabled" shortens but doesn't always remove the block. Keep thinking ON (default /
"adaptive") for hard reasoning, coding, and agentic loops.
When to use M3 vs M2.7
| Workload | Verdict |
|---|---|
| Clean chat / judgment / theory / JSON | ✅ M3 + reasoning_split:true (the new default) |
| Short tagging / classification, latency-sensitive | ✅ M3 + reasoning:"disabled", or stay on plain MiniMax-M2.7 |
| Vision (OCR, charts, screenshots) | ✅ M3 only — M2.7 is text-only; M3 reads images correctly |
| Structured JSON | ✅ M3 (response_format accepted) + reasoning_split + defensive parse |
| Long context (input up to ~1M) | ✅ input accepts to ~1M, but reliable retrieval ≤ ~256K (400K now misses); 1M prefill ~235 s |
| Raw math / QP / risk on realistic N | ❌ still route to Python (the M2.7 saturation guidance carries over) |
| Final deployable code | ⚠️ scaffold-only; sandbox-validate (unchanged from M2.7) |
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.
- 3d ago First seen · 100 lines · 162 tokens per session scan A 3ed7d524a02b
m3 is a skill published in the GitHub repository terrylica/cc-skills (62 stars, last pushed yesterday), licensed MIT. It adds 162 tokens to every session and 1,696 once invoked, about $0.0008 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 2026-09-05.
Other skills, from other repositories
ai-ml-development
AI and machine learning development with PyTorch, TensorFlow, and LLM integration. Use when building ML models, training pipelines, fine-tuning LLMs, or implementing AI features.
ai-policy-generator
AI governance policy creation for nonprofits and enterprises with frameworks, risk assessment, ethical guidelines, and compliance templates. Use when drafting AI usage policies, responsible AI frameworks, or organizational AI governance documents.
data-science
Data science and analytics expertise for statistical analysis, machine learning pipelines, data governance, business intelligence, predictive modeling, and analytics strategy. Use when building ML models, analyzing data, creating dashboards, or designing data architectures.
data-engineering
ETL/ELT pipelines, data warehousing (BigQuery, Snowflake, Redshift), stream processing (Kafka, Spark Streaming), orchestration (Airflow, Dagster, Prefect), dbt transformations, and data lake architecture. Use when building data pipelines, designing warehouse schemas, or implementing real-time data processing.
openrouter-models
Choose among the hosted models the shared router serves — discover current ids and prices from the public catalog, self-enforce a spend budget the router does not meter for you, respect the free-tier prompt-logging caveat, and request onboarding of a model the router does not serve yet.
llm-router-ops
Operate a self-hosted OpenAI-compatible LLM router/proxy (e.g. LiteLLM) in front of one or more backends — the minimal client-wiring block for every client type, the context-window advertisement gotcha, the env-vs-persisted-config gotcha, and why an unauthenticated health probe should 401, not 200. Use when wiring a…