Skill Github Daily Ops

Daily GitHub repo health check + safe Dependabot auto-merge. Outputs markdown report.

This skill mostly does what it claims (scan repos, report, and merge Dependabot PRs) but has several red flags you should address before using it with real credentials: - Expect to provide a GitHub PAT with repo write permissions. The registry doesn't declare this, but the scripts will fail without it. Use a token scoped as narrowly as possible (repo:status, pull_request, maybe repo if merging) and prefer an org-scoped machine account if possible. - The scripts use both GITHUB_TOKEN (node) and GH_TOKEN (bash) and will try to read ~/.github_token. Ensure you know which token will be used and avoid placing high-privilege tokens in a plaintext file unless you intend to. - The code relies on binaries not declared in the manifest: gh (GitHub CLI), python3, and git (for workspace push). Install these only if you trust the skill. - The included daily-ops.sh will attempt to push unpushed commits from a configured WORKSPACE (default ~/.openclaw/workspace). If you run that script, it may publish local commits you haven't reviewed. Either remove or audit the 'Push workspace commits' section before running, or set WORKSPACE to a safe path. - Consider running the scripts in readonly/report-only mode (--report) first, and test auto-merge behavior on a small set of non-critical repos. Review and possibly remove or sandbox any steps that write to remotes (merge, delete-branch, git push). Given the mismatches and the workspace-push side-effect, treat this skill as 'suspicious' until you fix the declared requirements, unify token handling, and remove or clearly control the workspace push behavior.

Type: OpenClaw Skill Name: skill-github-daily-ops Version: 1.0.1 The skill is classified as suspicious due to several risky capabilities, even though they align with the stated purpose of GitHub daily operations. Both `scripts/auto-merge.sh` and `scripts/daily-ops.sh` read the highly sensitive `GITHUB_TOKEN` from `~/.github_token`. Additionally, `scripts/daily-ops.sh` performs a `git push` operation on the local workspace, a powerful action that could lead to unauthorized code changes if the agent's environment or the workspace repository were compromised. While these actions are plausibly needed for the skill's functionality, they represent significant attack surfaces and potential vulnerabilities without clear malicious intent.