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 agentmods add commands/corvidlabs/agent-3md/create-specgit clone --depth 1 https://github.com/CorvidLabs/agent-3mdWhat 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 | $0.00030 | $0.00491 |
| Opus 5 | $0.00015 | $0.00246 |
| Sonnet 5 | $0.00006 | $0.00098 |
| Haiku 4.5 | $0.00003 | $0.00049 |
Grade A, and why
create-spec 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 2d 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.
What it actually says
Create a new spec-sync module spec.
Arguments: $ARGUMENTS
- Remove
--minimalwherever it appears and remember that minimal mode was requested. Preserve the complete remaining text; do not tokenize it yet. - Classify the complete remaining text:
- A bare module name — a short identifier like
auth-serviceorbilling, matching^[A-Za-z0-9][A-Za-z0-9._-]*$. Use it as-is. - A free-text feature description — a sentence or phrase describing
what to build, e.g.
"I want a feature that lets users export their data as CSV". In this case, invent a short, kebab-case module name that captures the idea (e.g.csv-export). If the right name is ambiguous, ask the user to confirm or rename it before continuing. Keep the full description at hand — you'll use it in step 5.
- A bare module name — a short identifier like
- If minimal mode was requested, run:
This creates a minimal spec only (no companion files).specsync new <module-name> - Otherwise (default), run:
This creates the spec, companion files (specsync scaffold <module-name>tasks.md,requirements.md,context.md,testing.md, anddesign.mdifcompanions.designis enabled), a registry entry, and auto-detects related source files. - Open the newly created
specs/<module-name>/<module-name>.spec.mdand fill in thePurpose,Requirements, andPublic APIsections. If a free-text description was given in step 2, use it directly to draft these sections — ask clarifying questions if it's underspecified, but do not leave the sections as unfilled placeholder text. Do the same forrequirements.md(acceptance criteria) andtasks.md(initial task breakdown), if present. - Run
specsync checkto confirm the new spec passes validation.
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.
- 2d ago First seen · 40 lines · 30 tokens per session scan A ce5e9d52554c
create-spec is a command published in the GitHub repository CorvidLabs/agent-3md (1 stars, last pushed 1mo ago), licensed MIT. It adds 30 tokens to every session and 491 once invoked, about $0.0002 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 2026-08-31.
Other commands, from other repositories
run-action
Execute a learned Maestro flow ("action") by name with optional -e KEY=VALUE parameters. Looks the flow up via packages/rn-dev-agent-core/dist/learned-actions.js (same inventory as /rn-dev-agent:list-learned-actions), then replays it via cdprunaction — auto-repair-aware orchestration with structured RunRecords (GH.
send-feedback
Send feedback or report a bug for the rn-dev-agent plugin. Collects sanitized environment context and creates a GitHub issue. No sensitive data (paths, secrets, PII) is transmitted.
end-session
Wrap up the work session with read-only evidence gathering, owner-classified documentation handoffs, gated issue sync, Kano refinement, and git hygiene checks.
test-feature
Test a React Native feature on the running simulator/emulator. Verifies UI, user flows, and internal state. Generates a persistent Maestro test file.
list-learned-actions
List persisted "learned actions" — feedback memories, Maestro flows, UI skeletons, and plugin commands that should be consulted before composing new device primitives. Wraps packages/rn-dev-agent-core/dist/learned-actions.js for programmatic discovery.
proof-capture
Capture PR-ready proof artifacts for a feature, with an attested fail-closed controller in strict mode.