Adaptive Trial Simulator
v0.1.0Design and simulate adaptive clinical trials with interim analyses, sample size re-estimation, and early stopping rules. Evaluate Type I error control, power...
⭐ 0· 114·0 current·0 all-time
byAIpoch@aipoch-ai
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
medium confidencePurpose & Capability
Name/description (adaptive clinical trial simulator) match the provided artifacts: SKILL.md documents simulation parameters and the included scripts/main.py implements sampling, alpha spending, conditional power, and sample‑size reestimation. No unrelated credentials, binaries, or config paths are requested.
Instruction Scope
SKILL.md tells the agent to run the included Python script with simulation flags; the instructions do not ask for reading unrelated files, environment variables, or network endpoints. The README claims 'No network access' and the visible portion of main.py contains only local numeric computations and file output. However the provided main.py content was truncated in this package listing, so the tail of the script (where unexpected behavior could appear) was not visible to me.
Install Mechanism
No install spec; it's instruction‑only plus a Python script. Dependencies are standard scientific packages (numpy, scipy, matplotlib) declared in requirements.txt — no remote/executable downloads or custom install scripts detected.
Credentials
The skill requests no environment variables, credentials, or config paths. Declared dependencies and runtime needs (Python 3.8+, NumPy/SciPy/Matplotlib) are proportionate to a statistical simulation tool.
Persistence & Privilege
The skill does not request always: true and does not modify agent/system configuration. It runs as a local script and writes simulation results (expected behaviour). No elevated persistence or cross‑skill config access is present in visible files.
Assessment
This skill appears to implement what it claims (a Monte Carlo adaptive trial simulator) and does not request secrets or system privileges. Before running: (1) inspect the entire scripts/main.py (search for network modules like requests/urllib/socket, subprocess/exec/eval, os.environ access, or obfuscated strings) since the provided listing was truncated; (2) run it in a sandbox or isolated environment first (container/VM); (3) pin or audit dependencies you install (numpy/scipy/matplotlib) to avoid supply‑chain surprises; (4) avoid feeding or storing real patient/PHI data with this tool unless you perform a privacy/security review; (5) for reproducibility and safety, run with small n_simulations first and check outputs. If you want, I can scan the rest of scripts/main.py (full file) for hidden network calls or credential access.Like a lobster shell, security has layers — review code before you run it.
latestvk972bm96mjjjsjqbbhhtjctscs8342nj
License
MIT-0
Free to use, modify, and redistribute. No attribution required.