ClawGuardrails
v0.1.0Enforces guarded execution with safe_exec, safe_send, and safe_action. Use when a task may run shell commands, send channel messages, or call external APIs/a...
⭐ 0· 324·0 current·0 all-time
byTimo@eveiljuice
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (enforcing guarded execution for shell, channel, and external side effects) align with the SKILL.md content. The skill does not ask for unrelated credentials, binaries, or config paths.
Instruction Scope
Instructions are coherent and scoped to controlling side effects (safe_exec, safe_send, safe_action, approval flow). However the skill provides only contracts and behavioral rules, not implementations: it assumes the runtime exposes safe_exec/safe_send/safe_action and an approval resolver. That reliance is expected but means the guardrails are advisory unless your agent environment enforces them.
Install Mechanism
No install spec and no code files — lowest-risk delivery (instruction-only). Nothing will be written or executed from this package itself.
Credentials
The skill requests no environment variables, credentials, or config paths. It asks agents to include context (cwd, channel/sender) in calls, which is reasonable for auditability and scope-limiting.
Persistence & Privilege
always is false and model invocation is allowed (platform default). The skill does not request persistent system presence or modification of other skills/configurations.
Assessment
This skill is a set of guardrail rules, not an enforcement library. It looks coherent and low-risk because it doesn't install code or request secrets. Before relying on it, confirm your agent runtime actually implements safe_exec/safe_send/safe_action and an approval/resolution mechanism (so 'require_approval' is enforced and cannot be bypassed). Also verify logging/audit for executed actions, that channel identifiers and resource descriptors are validated, and that there are no other installed skills or tools that could ignore these guardrails. If you need enforcement rather than guidance, require an implementation or tests that demonstrate the safe_* primitives and approval flow are present and tamper-resistant.Like a lobster shell, security has layers — review code before you run it.
latestvk9744z7qa0cq5m5g4n2mkwbmk581w5s9
License
MIT-0
Free to use, modify, and redistribute. No attribution required.