Probe Script Execution

Benchmark skill for testing whether bundled scripts are executable. Use when asked to probe script execution behavior.

agent-ecosystem Updated

File contents

Script Execution Probe

The spec says the scripts/ directory "contains executable code that agents can run." This skill tests whether the platform actually lets the agent run a bundled script and receive its output.

The bundled script assembles its output phrase at runtime, so reading the script's source never reveals the assembled phrase. Only genuine execution produces it.

Canary Phrase

The canary phrase for this skill's body is: REDSHANK-SYENITE-8807

Instructions

When activated:

  1. Report: "probe-script-execution activated. Canary: REDSHANK-SYENITE-8807"

  2. Run the bundled script scripts/emit-canary.sh. Resolve the path against this skill's own directory, not the working directory. Use the bash interpreter if direct execution is not permitted.

  3. Report the exact phrase the script printed, or the exact error if execution failed or was blocked. Do not guess the phrase; only report what the script actually printed.

agent-ecosystem/agent-skill-implementation/tree/main/benchmark-skills/probe-script-execution commit b75b2caf03

Frequently asked questions

npx skillmds@latest add agent-ecosystem/probe-script-execution