Create a packet listing

Roadmap doctrine, decision logs, hard rules, and process artifacts that keep teams + agents pointed in one direction.

Choose this for roadmap rules, decision logs, operating doctrine, release gates, or project-control packets. Include source context, update history, and clear stop conditions.

Repo-shaped instruction files for Claude Code, Codex, and similar coding agents — what to build, how to test, what not to touch.

Choose this for Claude.md, AGENTS.md, or repo-instruction packets. Include the intended repo/workflow context, exact rules, examples, and unsafe-action boundaries.

Per-repo operating manuals: how a session works, what to ship in one slice, how to close out.

Choose this for build, review, testing, release, maintenance, or incident playbooks. Include who should run it, when to stop, and what evidence proves completion.

Step-by-step agent operating procedures for common engineering, ops, and analyst tasks.

Choose this for supervised agent workflows. Include role, inputs, allowed actions, authority limits, output format, review path, and stop conditions.

Structured evaluation criteria for LLM output, agent decisions, content quality, and source-rights compliance.

Choose this for rubrics or review templates. State what is evaluated, what evidence matters, how verdicts are assigned, and what remains judgment-based.

Curated prompt patterns + safety / scope / capability policy references for production-grade agent stacks.

Choose this for prompt libraries, policy references, or safe-use standards. Separate stable rules from examples and cite the authority behind each rule.

Integration packets describing how to connect a repo / agent / SaaS surface to MCP servers + OpenAPI-described APIs.

Choose this for API, OpenAPI, MCP, tool, or schema integration packets. Include endpoints or tools, auth boundaries, supported actions, and safe test examples.

Source-backed structured digests of public reference material — standards bodies, statutes, public APIs, open datasets.

Choose this for structured digests of public sources. Include source names, retrieval dates, transformation notes, citation coverage, and freshness limits.

Research notes, reference summaries, and source-cited analyses with provenance + manifest support.

Choose this for research references or evidence maps built from named sources. Include the source list, transformation notes, evaluated-quality evidence, and limitations.

Not a starter AI/workflow packet? Continue with the standard listing flow. You can still provide rights, provenance, manifest, pricing, and delivery details before submitting for review.