Security

Supply chain security

repomatic implements most of the practices described in Astral’s Open Source Security at Astral post, baked into a drop-in setup that any maintainer can inherit by pointing their workflows at the reusable callers.

Astral practice	How repomatic covers it
Ban dangerous triggers (pull_request_target, workflow_run)	The lint-workflow-security job runs zizmor on every push: see .github/workflows/lint.yaml
Minimal workflow permissions	check_workflow_permissions parses every workflow file and warns when a custom-step workflow omits the top-level permissions key
Pinned actions	All uses: refs pinned to full commit SHAs (with the semver tag preserved as a trailing comment) via the sync-action-pins autofix job. check_sha_pinning_required warns when the repository’s sha_pinning_required setting is off, and the setup guide ships a fix: GitHub itself then refuses to run any workflow referencing an action by a mutable tag, closing the gap where a hand-edited workflow slips one past an inline-suppressed zizmor finding
No force-pushes to main	check_branch_ruleset_on_default verifies an active branch ruleset exists, and the setup guide walks users through creating one. check_classic_branch_protection reports any branch protection rule left beside that ruleset: GitHub still supports both and applies them together, so a rule surviving a migration splits the branch policy across two settings pages
Immutable release tags	check_immutable_releases verifies GitHub immutable releases is enabled, and the release workflow rewrites @main refs to @vX.Y.Z during freeze: see tagged workflow URLs
Dependency cooldowns	minimum-release-age shared cooldown for sync-tool-versions, sync-action-pins, and sync-workflow-pins; uv --exclude-newer for Python packages via sync-uv-lock, with a per-package escape hatch for CVE fixes: see minimum-release-age and cooldowns
Install-time cooldowns	Every workflow exports UV_EXCLUDE_NEWER and NPM_CONFIG_MIN_RELEASE_AGE at workflow level, so every uvx, uv pip install, uv tool install, npm install and npx in every job refuses a package published inside the window, transitive dependencies included. This covers the ad-hoc installs no pin or lockfile describes, including debugging steps and jobs added later: see install-time cooldown
Trusted Publishing	PyPI uploads via OIDC with no long-lived token. The publish-pypi job in each downstream caller workflow invokes the upstream publish-pypi composite action, which inherits the caller’s OIDC context. This sidesteps pypi/warehouse#11096, where reusable workflows mint an OIDC token whose job_workflow_ref does not match the downstream’s PyPI Trusted Publisher config. check_pypi_trusted_publisher reads PEP 740 provenance for the latest published file and warns when no bundle names this repo’s own release.yaml; the setup guide walks through the registration
Cryptographic attestations	Every binary and wheel is attested to the workflow run that built it via attest-build-provenance: see the Generate build attestations steps in .github/workflows/_release-engine.yaml
Checksums in installer scripts	The update-checksums CLI command regenerates SHA-256 checksums for every binary tool; invoked automatically by sync-tool-versions whenever a tool version is bumped
Fork PR approval policy	check_fork_pr_approval_policy warns when the policy is weaker than first_time_contributors, and the setup guide ships a pre-filled gh api one-liner to fix it

Warning

Known gap: multi-person release approval. Astral gates releases behind a dedicated GitHub deployment environment with required reviewers, so that a single compromised account cannot publish. repomatic does not enforce this, but if the repository has multiple maintainers, I recommend adding an environment: release key to the caller-side publish-pypi job (and to the upstream create-release job, if the caller exposes it) in the downstream workflow and configuring required reviewers on that environment in repo settings.

Important

One-time PyPI Trusted Publisher setup. Each downstream repository must register a Trusted Publisher entry on PyPI for its own caller workflow. The publisher config matches against the OIDC job_workflow_ref claim, which names the downstream’s workflow file (typically .github/workflows/release.yaml). Without this registration, the first PyPI upload after migration fails cleanly with a publisher mismatch error. See the PyPI Trusted Publishers documentation for the registration steps.

Third-party action minimization

Every third-party GitHub Action executes with access to GITHUB_TOKEN and repository secrets. Each action is a trust delegation: you depend on the maintainer’s security practices, their CI pipeline, and their transitive dependencies. A compromised action can steal secrets, inject code into builds, or tamper with releases.

repomatic has systematically eliminated 23 third-party actions since late 2025, replacing them with internal CLI commands, SHA-256-verified binary downloads, and runner built-in tools:

Removed action	Replacement	Strategy
calibreapp/image-actions	repomatic format-images	Internal CLI
crazy-max/ghaction-virustotal	repomatic scan-virustotal	Internal CLI
AndreasAugustin/actions-template-sync	repomatic init	Internal CLI
JasonEtco/is-sponsor-label-action	repomatic sponsor-label	Internal CLI
dessant/lock-threads	repomatic lock-threads	Internal CLI
actions/labeler	repomatic apply-labels	Internal CLI
github/issue-labeler	repomatic apply-labels	Internal CLI
peter-evans/create-pull-request	repomatic pr-sync	Internal CLI
lycheeverse/lychee-action	repomatic run lychee	Direct binary + SHA-256
crate-ci/typos	repomatic run typos	Direct binary + SHA-256
biomejs/setup-biome	repomatic run biome	Direct binary + SHA-256
gitleaks/gitleaks-action	repomatic run gitleaks	Direct binary + SHA-256
julb/action-manage-label	repomatic run labelmaker	Direct binary + SHA-256
taiki-e/install-action	Direct curl + checksum	Direct binary + SHA-256
softprops/action-gh-release	gh release create	Runner built-in
actions/github-script	Bash + gh CLI	Runner built-in
crazy-max/ghaction-dump-context	Bash + runner built-ins	Runner built-in
actions-rust-lang/setup-rust-toolchain	Runner built-in Rust	Runner built-in
actions/setup-python	astral-sh/setup-uv	Consolidated
peaceiris/actions-gh-pages	actions/deploy-pages	First-party replacement
codecov/codecov-action	None (integration dropped)	Removed entirely
codecov/test-results-action	None (feature dropped)	Removed entirely
GitHubSecurityLab/actions-permissions	Explicit permissions: key	Removed entirely

The only remaining third-party action (1 of 8 total) is astral-sh/setup-uv, which installs the toolchain every other job runs through.

That makes its own pin do double duty: it fixes the action’s code, and it fixes which uv builds arrive verified against a pinned hash, since setup-uv checks a download against the table its release bundles. A build missing from that table installs with no verification before v10.1.0, and against a hash fetched at run time from v10.1.0 on. sync-workflow-pins therefore never walks the uv pin past that table, and steps it back onto the table when it finds a pin already sitting above one: the action bump that would otherwise repair it does not exist while the newest setup-uv is the one already pinned. lint-repo reports the same condition.

Every other uses: in the tree is GitHub’s own: actions/checkout, actions/cache, actions/upload-artifact, actions/download-artifact, actions/attest, actions/upload-pages-artifact and actions/deploy-pages. All but actions/checkout speak the Actions cache, artifact, attestation and Pages backplanes, and are kept deliberately: those are internal service protocols GitHub reserves the right to change, so a reimplementation would break silently rather than loudly.

Issue and pull-request labelling

repomatic apply-labels covers what the two retired labeller actions did: file globs over a pull request’s changed paths (actions/labeler) and content patterns over its title and body (github/issue-labeler). One job runs it on every issue and pull request opened; repomatic.labels holds the matcher, tests/test_labels.py pins the semantics, including the minimatch glob dialect.

The supply-chain gain is the usual one, and modest: both actions were SHA-pinned with cooldown-gated bumps. The concrete win is that the rules stopped being files. Each job used to run repomatic init labels to stage a YAML config purely so its action could read it back out of the checkout, and the actions’ schemas dictated the config’s shape all the way up into [tool.repomatic.labels]. Owning the matcher retired the staging step, the two bundled YAML files, the serializers that produced them, and the schema itself: a rule is now one label mapped to the keywords or globs that apply it, with the defaults living in repomatic.labels.DEFAULT_CONTENT_RULES and DEFAULT_FILE_RULES as plain Python data.

It also fixed a silent bug the old shape could only warn about. github/issue-labeler AND-joined a label’s patterns, so the obvious “any of these keywords” rule matched nothing; any pattern matching now applies the label, and keywords are matched case-insensitively on word boundaries instead of each rule hand-rolling /\bword\b/i.

Pull-request creation

repomatic pr-sync opens, refreshes and retires every pull request CI produces, across all 22 call sites that used to run peter-evans/create-pull-request. It is a deliberate port of that action’s algorithm rather than a fresh design, and repomatic.github.pr records the mapping.

That action was the last third-party one entrusted with REPOMATIC_PAT, a token carrying workflow-write scope, which made it the largest remaining trust delegation on this page. It was nonetheless SHA-pinned with bumps gated on minimum-release-age by sync-action-pins, so the supply-chain gain is real but bounded: a compromised release could not have reached a build until it had been public for the cooldown window and a maintainer had merged the bump.

The concrete win is elsewhere. The action’s cleanup only ran when the action ran, and a job behind an if: gate skips its own steps, leaving a branch and pull request nobody retires: repomatic close-stale-bump-pr exists to patch exactly that hole. pr-sync decides internally instead, so one command covers the create, update, no-op and close cases.

Owning both ends also collapsed the plumbing: each job used to render its body with repomatic pr-body, squeeze it through a $GITHUB_OUTPUT step output (a 32 KiB ceiling), and relay it into the action via env:. pr-sync --template X renders internally, so the relay, its size ceiling, and the per-step env: blocks are gone, and the branch, labels and draft state derive from the template’s own frontmatter.

A conformance test holds this in place: test_pull_requests_are_opened_by_the_cli_not_an_action fails when any workflow reaches for the action again. Without --base on a detached HEAD, pr-sync reads the repository’s default branch from the CI event payload, and fails when the payload names none.

Replacement strategies, ordered from most to least isolated:

1. Internal CLI: the operation runs inside repomatic Python code with no external process.

2. Direct binary download: checksummed binary fetched from a GitHub release URL, no action code path involved.

3. Runner built-in: uses tools pre-installed on the GitHub Actions runner (gh, Rust toolchain).

4. First-party replacement: swaps a community action for an official actions/* equivalent maintained by GitHub.

Ruff consolidation

Ten separate Python linters and formatters, two of them mdformat plugins, have been absorbed into ruff, eliminating ten runtime or dev dependencies:

Removed tool	What it did	Replaced
pylint	Static analysis and linting	Feb 2023
pydocstyle	Docstring convention enforcement	Feb 2023
pycln	Unused import removal	Feb 2023
pyupgrade	Python syntax modernization	Feb 2023
isort	Import sorting	Feb 2023
black	Code formatting	Sep 2023
docformatter	Docstring formatting	Jan 2024
mdformat-black	Python formatting in Markdown code blocks, as an mdformat plugin	Aug 2024
blacken-docs	Python formatting in Markdown code blocks	Feb 2026
mdformat-ruff	Same as mdformat-black, through a second ruff pinned inside the mdformat environment	Aug 2026

autopep8 is the only legacy formatter still in use: it handles long-line comment wrapping that ruff does not yet cover (astral-sh/ruff#7414).

uv consolidation

Five separate packaging and install tools have been absorbed into uv, which now handles dependency management, builds, publishing, auditing, and Python version installation:

Removed tool	What it did	Replaced
poetry	Dependency management, lock files, virtual environments	Jun 2024
build / python -m build	Package building (wheels and sdists)	Sep 2024
twine	PyPI uploads	Jan 2025
check-wheel-contents	Wheel validation	Jan 2025
pip-audit	Vulnerability scanning	Mar 2026

uv also consolidated command-line usage that previously required separate tools: pip install became uv pip install / uv sync, pipx became uvx, and actions/setup-python was replaced by astral-sh/setup-uv (counted in the action minimization table above).

Two other Python packages were eliminated outside the ruff/uv consolidations: pipdeptree (replaced by an internal dep-graph implementation) and gitignore-parser (replaced by py-walk).

Permissions and token

Several workflows need a REPOMATIC_PAT secret to create PRs that modify files in .github/workflows/ and to trigger downstream workflows. Without it, those jobs silently fall back to the default GITHUB_TOKEN, which lacks the required permissions.

After your first push, the setup-guide job automatically opens an issue with step-by-step instructions to create and configure the token.

Concurrency and cancellation

All workflows use a concurrency directive to prevent redundant runs and save CI resources. When a new commit is pushed, any in-progress workflow runs for the same branch or PR are automatically cancelled.

Workflows are grouped by:

• Pull requests: {workflow-name}-{pr-number} — Multiple commits to the same PR cancel previous runs

• Branch pushes: {workflow-name}-{branch-ref} — Multiple pushes to the same branch cancel previous runs

release.yaml uses a stronger protection: release commits get a unique concurrency group based on the commit SHA, so they can never be cancelled. This ensures tagging, PyPI publishing, and GitHub release creation complete successfully.

Additionally, cancel-runs.yaml actively cancels in-progress and queued runs when a PR is closed. This complements passive concurrency groups, which only trigger cancellation when a new run enters the same group — closing a PR doesn’t produce such an event.

Tip

For implementation details on how concurrency groups are computed and why release.yaml needs special handling, see the repomatic.github.actions module docstring.

AV false-positive submissions

Compiled Python binaries (built with Nuitka --onefile) are frequently flagged as malicious by heuristic AV engines. The onefile packaging technique (self-extracting archive with embedded Python runtime) triggers generic “packed/suspicious” signatures. This is a known issue across the Nuitka ecosystem.

The scan-virustotal job in _release-engine.yaml uploads all compiled binaries to VirusTotal on every release. This seeds AV vendor databases to reduce false positive rates for downstream distributors (Chocolatey, Scoop, etc.). Each release’s flagged / total snapshot is recorded in docs/assets/virustotal-scans.csv and rendered, along with the full catalog of released binaries and their analysis links, on the binaries page. Detection counts are deliberately kept out of GitHub release notes, where they read as a malware verdict without context (see kdeldycke/meta-package-manager#1911).

When a release is flagged, the /av-false-positive skill generates per-vendor submission files with pre-written text and form field mappings. The vendor details below document the process for manual reference.

Why binaries get flagged

Nuitka --onefile creates a self-extracting archive that decompresses an embedded Python runtime to a temporary directory and executes it at launch. This “drop and execute from temp” pattern is behaviorally identical to trojan droppers, which triggers heuristic and ML-based detections. Two more factors compound it: Nuitka is popular with malware authors for source code protection, which poisons AV heuristics for all Nuitka-compiled binaries, and Microsoft has gone as far as suspending an Artifact Signing account over Nuitka onefile binaries.

The detection profile is consistent across projects: Linux binaries scan clean, macOS ones pick up the occasional ML false positive, Windows ARM64 stays low (fewer ARM64 heuristics in AV engines), and Windows x64 attracts the bulk of the detections through generic signatures like Gen:Variant.Application.tedy (BitDefender family), Trojan:Win32/Sabsik (Microsoft), Python/Packed.Nuitka.AL (ESET), and various ML classifiers. Pure-Python .whl and .tar.gz distributions scan clean.

The Nuitka project tracks the situation in Nuitka/Nuitka#2685, Nuitka/Nuitka#2495, Nuitka/Nuitka#2757, and Nuitka/Nuitka#3842.

Vendor portals

Vendor	Engines covered	Portal	Format	Turnaround
Microsoft	Microsoft	WDSI file submission	One file per form, 1900 char limit on additional info	Fastest
BitDefender	BitDefender, ALYac, Arcabit, Emsisoft, GData, MicroWorld-eScan, VIPRE	bitdefender.com/submit	One file per form, screenshot mandatory	Fast
ESET	ESET-NOD32	Email to samples@eset.com	Single email, password-protected ZIP (infected), ~24 MB limit	Reliable
Symantec	Symantec	symsubmit.symantec.com	Hash submission only (no .exe/.bin upload), one hash per form, 5000 char limit	3-7 business days
Avast/AVG	Avast, AVG	avast.com/submit-a-sample	One file per form, shared engine	Medium
Sophos	Sophos	sophos.com filesubmission	One file per form, 25 MB max per submission	Up to 15 business days

Complete directories of vendor false-positive contacts are maintained by VirusTotal and False-Positive-Center.

Submission priority

Submit in this order to maximize impact:

1. Microsoft: most influential engine. ML detections (Sabsik, Wacatac) have the broadest downstream effect.

2. BitDefender: powers ~6 downstream vendor engines. Highest detection-removal-per-submission ratio.

3. ESET: email-based channel with no portal dependency. The most reliable submission path.

4. Symantec: ML detections (ML.Attribute.*) may take longer to process.

5. Avast/AVG: shared engine, so one submission covers both.

6. Sophos: PUA detections require justification of the software’s legitimate purpose.

Submission content

Every false-positive submission should include:

• The binary’s VirusTotal report link.

• VirusTotal links for the clean .whl and .tar.gz source distributions (as comparison evidence).

• The GitHub release link and direct download URL for the binary.

• Project homepage and PyPI URL.

• License from pyproject.toml.

• Reference to any prior false-positive issue in the repository.

All submission text should mention that the binary is compiled with Nuitka --onefile from an open-source project.

Known portal issues

• Microsoft: CORS errors or stuck progress modals during upload (auth session expiring). Workaround: sign out, clear cookies for microsoft.com, sign back in, submit immediately.

• BitDefender: form sometimes returns “Your request could not be registered!” with no details. Retry later.

• Avast: form sometimes returns “An internal error occurred while sending the form.” Retry later.

Long-term mitigations

False-positive submissions are a per-release moving target. The structural fixes:

• Code signing with an EV certificate would reduce heuristic detections across the board, especially from Microsoft and Symantec ML models.

• Switching from --onefile to --standalone would eliminate the self-extracting pattern entirely, at the cost of distributing a directory instead of a single .exe.

• Nuitka Commercial claims proprietary AV-mitigation techniques but offers no guarantees.

The code behind this page is documented on the repomatic.release.virustotal, repomatic.release.checksums and repomatic.release.binary pages.