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 mehrad-dm/mastermind --skill helpgit clone --depth 1 https://github.com/mehrad-dm/mastermindWrote 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/mehrad-dm/mastermind/help)<a href="https://agentmods.dev/skills/mehrad-dm/mastermind/help"><img src="https://agentmods.dev/badge/skills/mehrad-dm/mastermind/help/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/mehrad-dm/mastermind/help"><img src="https://agentmods.dev/badge/skills/mehrad-dm/mastermind/help.svg" alt="Reviewed on agentmods" width="80" 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 Agent Snooping · line 98 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.
- medium Agent Snooping · line 99 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 | $0.00061 | $0.02419 |
| Opus 5.5 | $0.00024 | $0.00968 |
| Sonnet 5.5 | $0.00012 | $0.00484 |
| Haiku 4.5 | $0.00006 | $0.00242 |
Grade A, and why
help 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 15d 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 — 110 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Help: what MasterMind can do, and how to use it
Present this as a clear, friendly menu (adapt length to the ask). Open with the brand header, then lead with the one thing that matters most, then the tables. Keep it scannable. Header:
🧠 MasterMind: a genius-builder brain for your AI
33 skills · 4 agents · Claude Code · Cursor · Codex
Lead with what fits this project
Before the full menu, name up to three skills that would earn their keep here right now, each with the reason it applies: read from what's actually in front of you (the stack, the current task, what the last few turns were about, whether a field pack exists yet):
Most useful here right now
· init: no field pack yet for this Django project; one-time setup, then I'm sharper on it
· quarantine: there's client data in this repo; this keeps it out of git
· handoff: this session is long; a handoff survives the reset
A full menu is a catalogue; three fitting picks are advice. Then give the tables below: the catalogue still matters, because the point is that the user learns what they already own.
The one rule: you don't run commands: you just talk
MasterMind works by recognizing intent. Say what you want in plain language and it applies the right
skill or agent itself, then shows a short note that it did (e.g. 🧠 MasterMind ▸ building this, will verify before handoff, with └ build · design → implement → verify beneath it). The /name calls below are an optional shortcut: a way to force a specific skill, not
something you need to memorize. Every skill is Auto (the model can invoke it); you can also call any
of them by hand.
Skills: the workflows (all Auto-invocable)
| Skill | What it does | Fires automatically when… | Call it by hand |
|---|---|---|---|
| init | set MasterMind up for this project: detect the stack (or ask what you're building), load the field pack | your first real task in a project it isn't set up for | "init" · "set up MasterMind" |
| build | design → implement → verify → review → capture, end to end | you ask to build/add/implement a non-trivial feature | "build a search box for the header" |
| debug | structured, evidence-first root-cause debugging | a bug is hard, weird, or keeps coming back | "debug why the list doesn't refresh" |
| performance | measure → find the real bottleneck → fix the biggest → verify | something's slow or janky ("why is this slow?") | "why is this page so slow?" |
| qa | prove a change works by driving it for real (+ optional tests) | you finish something / ask to test it | "qa this" · "write tests for it" |
| report | a shareable write-up of a cycle: changed files, decisions, verification, verdict (MD default, HTML on request) | you ask for a report; or end of build/qa if you turned it on (off by default) | "give me a report of this" · "reports on" |
| interview | turn a fuzzy ask into a crisp, buildable spec; resolve a written spec's open questions | the request is ambiguous, or a spec has open questions | "interview me about the checkout flow first" |
| assess | decide whether an idea deserves building: go, clarify or stop, with evidence | an idea nobody has said yes to yet | "is this worth building?" |
| constitution | write and version the rules every feature in this project must obey | you state a non-negotiable, or start spec-first work | "these are our non-negotiables" |
| living-docs | keep the product, business and tech documents current, each fact written once | docs are missing or stale, or a feature just finished | "write down how this product works" |
| specify | write a feature spec: what and why, prioritized stories, traceable requirements | a feature spans sessions, or touches money, auth or data | "write the spec for invitations" |
| blueprint | turn a spec into a technical plan, checked against the constitution | a spec has no plan yet | "plan this spec" |
| breakdown | cut the plan into tasks, grouped by user story so each ships alone | a plan has no task list yet | "break this into tasks" |
| checklist | test one concern's requirements (security, UX, API) before building | a sensitive feature heads into planning | "security checklist for this spec" |
| analyze | check spec, plan and tasks agree before any code, read-only | you're about to build from a task list | "is this ready to build?" |
| converge | check the built code against everything the feature promised; append the gaps | every box is ticked, or you ask "is it really done?" | "check the code against the spec" |
| route | load only the files/skills a task actually needs | a task is non-trivial or you're unsure where to start | "route this" |
| learn | get up to speed fast on an unfamiliar stack/API, verify assumptions | the tech is new, fast-moving, or you're guessing | "learn how the Stripe API handles refunds" |
| prototype | build a throwaway prototype to learn, then rebuild properly | the right approach is genuinely unknown | "throw together a quick prototype of this" |
| signature | capture your team's real conventions as durable rules it follows | you ask it to match your codebase / keep correcting the same thing | "make it match our patterns" |
| persona | write in the documented public style of a named engineer, cited | you ask for someone's style ("write like X") | "write this like Kent C. Dodds" |
| explain | write AI-friendly usage docs beside an internal package | an internal package has little/no docs (asks first) | "explain this package" |
| prompt | sharpen a vague prompt into a precise, AI-ready one | you ask to optimize/fix a prompt for an AI | "improve this prompt" |
| quarantine | set up a private, gitignored quarantine for confidential data | before capturing project/client-specific notes | "keep this client's data private" |
| levelup | improve MasterMind's own knowledge (lessons, curriculum, new field) | after a correction/review, or to refresh standards | "levelup: remember this" |
| double-check | interrogate a claim before handing it over: adversarial, reviewer never sees your conclusion | you're about to say "it works", or a review came back suspiciously clean | "are you sure? double-check it" |
| deprecate | remove/migrate/retire safely: expand → migrate → contract | deleting a feature, column, flag, or endpoint | "can we delete this?" |
| roadmap | a durable decision map for multi-week work | work spanning many sessions; "why did we decide X?" | "keep a roadmap for this project" |
| lint | check MasterMind's own files for bloat, repeated rules, contradictions | before a release, or the brain feels bloated | "lint the brain" |
| clarify | re-pitch an answer that did not land, in plain words or as a picture | you say "I did not follow that", "simpler", "you lost me" | "say that simply" |
| deepen | survey an area and rank where the design costs most, you pick one | "this is getting messy", "why is this so hard to change?" | "where is this hurting?" |
| handoff | write a concise handoff so work survives a reset/teammate | before a long-task /clear, pausing, or handing off |
"write a handoff" |
| help | show this menu, what every skill/agent does and when it fires | you ask "what can you do?" / "help" / "how do I use this" | "what can you do?" |
What ships with it
1 file 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.
- 15d ago Changed · +10 lines · +61 tokens per session 24f2f3bb541f
- 1mo ago First seen · 100 lines · 0 tokens per session scan A 9b5ca8160754
help is a skill published in the GitHub repository mehrad-dm/mastermind (23 stars, last pushed 6d ago), licensed MIT. It adds 61 tokens to every session and 2,419 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 2026-08-30.
Other skills, from other repositories
st-full-workflow
Use when the user asks to run the complete end-to-end Strikethroo workflow for a work order in one shot in this repository — triggers include full workflow, end-to-end, plan and execute, do everything, run the whole strikethroo workflow. Do not use when the user wants only one stage (create a plan, generate tasks, or…
code-sec-review
Static security review that hunts exploitable vulnerabilities, prioritizing unauthenticated and cross-tenant attack paths.
agent-decision
Deterministic self-evaluation rubric for decision escalations — scored every run using the FRAME framework.
code-quality-review
Reviews code and plans against the project's coding rules.
review-loop
Two-mode review loop — single reviewer per epic, three reviewers for full branch.
setup-new-acpx-key
Configure the API key (or other credential) for a registered ACPX agentid (claude/codex/cursor/gemini/qwen/copilot/pi/droid/iflow/kilocode/kimi/kiro/opencode/tlamatini) end-to-end across data.keys, config.json (top-level + acpx.agents. .env), regensecrets.py, and verification via acpdoctor.