OpenAgentd / scripts
AGENTS.md217 starsChanged 28 days agoApache-2.0View on GitHub
This subtree owns repository validation, release/version maintenance, icon generation, updater helpers, and code-health analysis. - validate_docs.py: Markdown links/frontmatter and documented Make target contracts. - codehealth/: stdlib analyzer for Python/TypeScript size, complexity, coupling, and import cycles. make health and make health-json analyze the web frontend only; --lang python analyzes scripts/. - brand/: the logo's source of truth. brand.ts (its own Bun package, with Schibsted Grotesk and its OFL license in fonts/) builds the Seal and the wordmark from the font and…
- Reads credentials
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 31,700 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository8 other files it gives its agents.
- .agents/AGENTS.mdAGENTS.md
- AGENTS.mdAGENTS.md
- appv3/AGENTS.mdAGENTS.md
- desktop/AGENTS.mdAGENTS.md
- documents/AGENTS.mdAGENTS.md
- mobile/AGENTS.mdAGENTS.md
- web/AGENTS.mdAGENTS.md
- guidelinesSkill
The file
We split this AGENTS.md into the rules it gives an agent
2 rules. About 1 is common, 0 are distinctive, and 1 is very specific to this project.
≈1
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 51
Do not run version bump, signing-key, manifest, publish, or release helpers merely to verify documentation.
- Common: line 54
Never embed signing material, tokens, or machine-local paths.
≈ “Common” is an estimate: at least 1 in 5 of the 40,246 AGENTS.md files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Maintainer Scripts Guide
This subtree owns repository validation, release/version maintenance, icon generation, updater helpers, and code-health analysis.
Ownership
validate_docs.py: Markdown links/frontmatter and documented Make target contracts.codehealth/: stdlib analyzer for Python/TypeScript size, complexity, coupling, and import cycles.make healthandmake health-jsonanalyze the web frontend only;--lang pythonanalyzesscripts/.brand/: the logo's source of truth.brand.ts(its own Bun package, with Schibsted Grotesk and its OFL license infonts/) builds the Seal and the wordmark from the font and writesweb/src/lib/brand/logo-data.ts, the logo files and favicons underdocuments/assets/brand/, the platform icon masters,web/public/favicons, and the boot loader's inline Seal inweb/index.html. Runbun install --cwd scripts/brand && bun scripts/brand/brand.ts, thenpython3 scripts/generate_icons.py --force. openagentd.dev copies the logo from here; never redraw it by hand.generate_icons.py: turns the brand masters into every native icon (desktop.icns/.ico/PNGs and tray template, iOS and Android launcher icons) throughcargo tauri icon.make_updater_manifest.pyandgenerate_updater_keys.sh: desktop updater metadata and local key setup.release_version.sh,bump_version.sh,check_version_consistency.sh, andrelease_commits_since_last_tag.sh: synchronized release metadata and release-note inputs. The[workspace.package]version inappv3/Cargo.tomlis the release version.e2e_cli_install.shandfake_release_server.py: CLI install and self-update end-to-end check used by the appv3 workflow.readme-shots/: regenerates the README screenshots and banner from a throwaway instance.run.sh upstarts an isolated backend and web UI (HOME and every data directory under/tmp/oad-demo, invented demo repos fromdemo-repos.sh),run.sh stageruns real agent sessions (stage.py, a few cents of model calls using keys from.env),run.sh capture [scene…]takes light and dark shots intodocuments/assets/readme/(capture.mjs), andbanner.mjsrendersdocuments/assets/brand/banner/banner.html. Never point it at a real data directory: screenshots must not show user sessions or paths.tests/: pytest checks for these scripts and for installer, native-shell config, and release-workflow contracts; run them withmake verify-scripts.
The repository has no Python project. Scripts run with python3 and the
standard library (plus Pillow for generate_icons.py); make verify-scripts
installs the test dependencies into the local, git-ignored .venv. Keep scripts
non-interactive by default, repository-root-relative, and portable across
supported platforms.
Safety and generated outputs
- Do not run version bump, signing-key, manifest, publish, or release helpers
merely to verify documentation. Inspect
--helpor source and use the smallest non-mutating path. - Never embed signing material, tokens, or machine-local paths. Updater private keys remain outside the repository.
- Sidecar bundles, Cargo targets, web distributions, and generated native platform trees are build output. Change source inputs and rerun their owning script/Make target.
- Version changes must use the release workflow so the appv3 workspace, web,
desktop, mobile, Tauri configs, lockfiles, and feature-catalogue metadata
stay in sync.
make verify-versionis the gate.
Checks
Choose the focused safe command, then the owning repository target:
python3 scripts/validate_docs.py
python3 scripts/make_updater_manifest.py --help
python3 -m scripts.codehealth --help
make verify-scripts
make verify-docs
make verify-versionFor sidecar changes, run make -C desktop sidecar when feasible. For icon,
release, updater, or packaging changes, report any platform/signing step that
could not be exercised locally.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

