Wiki Lint
v0.1.0Lint and health-check a persistent markdown knowledge wiki. Use when a user wants the agent to inspect a wiki for contradictions, stale claims, missing cross...
⭐ 0· 63·0 current·0 all-time
by@teki-ai
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name and description match the instructions: the skill inspects markdown wikis (index.md, log.md, sample pages) and produces reports or fixes. There are no unrelated environment variables, binaries, or install steps requested.
Instruction Scope
SKILL.md explicitly instructs the agent to inspect wiki files and, if asked, implement fixes. That is coherent with linting, but the instructions give the agent broad discretion over which pages to read and what edits to make; the skill does not constrain scope, require user confirmation before writes, or declare where edits should be committed (filesystem vs. git).
Install Mechanism
No install spec and no code files — instruction-only. This minimizes risk from downloaded/installed code.
Credentials
No environment variables, credentials, or config paths are requested, which is proportional for a lint-only skill. Note that actual operation will require file-system or repository access provided by the host environment; the skill does not request network or service tokens.
Persistence & Privilege
always:false and default agent invocation are appropriate. The skill does not request persistent system modification rights or to change other skills' configs. However, because it may implement edits, granting write/commit privileges to the running agent increases risk and should be controlled by the user.
Assessment
This skill appears to do what it says: review markdown wiki files and report or apply fixes. It does not request any secrets or installs, which is good. However, the runtime instructions let the agent read arbitrary wiki files and — if you ask it to — make edits. Before enabling or invoking the skill: (1) limit its access to only the repository or path you want it to inspect (avoid giving access to directories that may contain secrets); (2) prefer a read-only run first and review the generated report before allowing edits; (3) if you allow edits, require manual review or run in a sandbox branch (git) so you can inspect and approve commits; and (4) do not grant unnecessary network or credential access (e.g., CI tokens, cloud keys) just to run this lint. If you need the agent to make changes automatically, consider adding explicit constraints in the prompt or an approval step so edits are scoped and auditable.Like a lobster shell, security has layers — review code before you run it.
knowledge-managementvk978qt4gkc29s2t96fxva1n5h584qb6qlatestvk978qt4gkc29s2t96fxva1n5h584qb6qmarkdownvk978qt4gkc29s2t96fxva1n5h584qb6qobsidianvk978qt4gkc29s2t96fxva1n5h584qb6qwikivk978qt4gkc29s2t96fxva1n5h584qb6q
License
MIT-0
Free to use, modify, and redistribute. No attribution required.