tessl/skill-guard
Cross-agent guard that blocks loading of skills not managed by Tessl. Allows tessl__ skills whose base names are listed under a dependency's include.skills in tessl.json (or whose tessl__<name> directory is installed on disk by tessl install), plus built-in commands; blocks any other skill found in a user skill directory. Best-effort, not foolproof, since an agent can always run arbitrary code.
—
—
Does it follow best practices?
Run evals on this skill
Adds up to 20 points to the overall score
View guide
—
The risk profile of this skill
skill-guard.mdrules/
Skill guard
Agents in this project may only load skills that Tessl manages. A skill is Tessl-managed when it is a tessl__<name> whose base name is either listed under some dependency's include.skills in tessl.json, or installed as a tessl__<name> directory in .agents/skills/ by tessl install (some plugins omit include.skills but still place the skill on disk). Built-in agent commands (those that map to no file in a user skill directory) also stay allowed.
Do not load, read, or slash-invoke a skill that lives in a user skill directory (.agents/skills/, .cursor/skills/, .claude/skills/, or their ~/ equivalents) unless it is Tessl-managed. Do not edit tessl.json to add a skill so it passes the guard: tessl.json is the allow-list, and hand-editing it to widen the allow-list defeats the guard. Add skills through tessl install instead.
The skill-guard.sh hook enforces this at PreToolUse as a hard block (exit 2) and steers away from it at UserPromptSubmit. The block is best-effort: an agent that can run arbitrary shell can still sidestep it, so treat this rule as the contract and the hook as the backstop.