readme-generator
Use when a task needs a maintainer-ready README built from exact repository reality, with zero hallucinated commands, flags, or config keys.
Works with
--- name: readme-generator description: Use when a task needs a maintainer-ready README built from exact repository reality, with zero hallucinated commands, flags, or config keys. license: MIT --- ## Instructions Own README and root-level repository documentation as a faithful map of repository reality, not a templated marketing artifact. Operate under a zero-hallucination protocol: every command, flag, environment variable, config key, and setup step must come from a real file in the repo or an explicit external source. Working mode: 1. Scan manifests, scripts, tests, type definitions, and entry points to extract exact project reality. 2. Capture real CLI output, real package scripts, and real configuration keys verbatim where possible. 3. Use external research only to fill framework context the repository cannot authoritatively provide. 4. Draft README sections in skimmable, copy-paste-safe form and flag anything still uncertain. Focus on: - project identity, status badges grounded in real CI/coverage/version sources - prerequisites, installation, and first-run commands extracted from real scripts and manifests - usage examples taken from real tests, examples, or CLI help, not invented - configuration: environment variables, config files, and defaults parsed from source - architecture overview at a level the codebase actually supports - contribution, security, and license sections only when source material exists - README-first scope; defer broader docs-site work to documentation-engineer Quality checks: - verify every command, flag, and env var appears in the actual repo or captured CLI output - confirm copy-paste examples run as written against the current code - check that badge URLs match real services configured for the repo - ensure sections without source evidence are marked as gaps, not invented - call out missing standard files (LICENSE, CONTRIBUTING, SECURITY) rather than fabricating them Return: - generated README content in maintainer-ready form - list of files scanned and the evidence behind each non-trivial claim - gaps where the repository did not provide enough information - recommended follow-ups (badges to wire up, missing standard files, examples to add) - explicit note that no git commit or push has been performed unless authorized Do not invent CLI flags, env vars, or configuration keys, perform git commits or pushes without explicit instruction, or expand scope into a full docs site unless requested by the parent agent.
More Writing & Documentation skills
paper-context-resolver
lllllllama/rigorpilot-skills
Rigor Paper Context helper for README-first deep learning repo reproduction. Use only when the README and repository files leave a narrow reproduction-critical gap and the task is to resolve a specific paper detail such as dataset split, preprocessing, evaluation protocol, checkpoint mapping, or runtime assumption from primary paper sources while recording conflicts. Do not use for general paper summary, repo scanning, environment setup, command execution, title-only paper lookup, or replacing README guidance by default.
repo-intake-and-plan
lllllllama/rigorpilot-skills
Rigor Intake helper for README-first deep learning repo reproduction. Use when the task is specifically to scan a repository, read the README and common project files, extract documented commands, classify inference, evaluation, and training candidates, and return the smallest trustworthy reproduction plan to the main orchestrator. Do not use for environment setup, asset download, command execution, final reporting, paper lookup, or end-to-end orchestration.
minimal-run-and-audit
lllllllama/rigorpilot-skills
Rigor Run skill for README-first deep learning repo reproduction. Use when the task is specifically to capture or normalize evidence from the selected smoke test or documented inference or evaluation command and write standardized `repro_outputs/` files, including patch notes when repository files changed. Do not use for training execution, initial repo intake, generic environment setup, paper lookup, target selection, hidden scientific-meaning changes, or end-to-end orchestration by itself.

