Ib Summarizer
v0.1.0Summarize core safety information from Investigator's Brochures for clinical researchers
⭐ 0· 67·0 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, SKILL.md, and scripts/main.py are consistent: the tool reads local PDF/DOCX/TXT, extracts safety-related sections, and builds summaries. Declared optional AI libraries (openai/anthropic) are listed as optional for 'AI-enhanced extraction' which is plausible but not required for core functionality.
Instruction Scope
SKILL.md instructs running the local Python script on an input file and writing outputs; that scope is appropriate. The checklist asks for path validation and sandboxing, but the provided code does not advertise explicit sandbox/OCR tooling (it punts OCR to the user). The script appears to read arbitrary files specified by the user — expected for this tool, but you should validate input-file handling and output paths before running on sensitive systems.
Install Mechanism
No install spec; dependencies are standard Python libraries installed via requirements.txt. There are no remote download URLs or extracted archives. Minor oddities: requirements.txt lists both 'docx' and 'python-docx' and includes 'dataclasses' although Python 3.8+ already provides dataclasses; these are bookkeeping issues rather than security risks.
Credentials
The skill requires no environment variables, no credentials, and no special config paths. The optional mention of openai/anthropic would introduce API keys if used, but those are optional and not required by the core script.
Persistence & Privilege
The skill is not always-enabled and does not request persistent system privileges or modify other skills. It operates as an invoked local script and does not require autonomous persistent presence.
Assessment
This skill appears coherent for extracting safety info from IB documents, but before installing/running: (1) inspect the full scripts/main.py for any unshown code paths (the provided file was partially truncated in the prompt); (2) run it on non-sensitive sample documents first to verify behavior; (3) run in a sandbox or isolated environment because it reads arbitrary files you point it at; (4) if you plan to enable the optional OpenAI/Anthropic features, be aware that you will need to provide API keys and that introduces network calls and data-leak considerations; (5) validate that output summaries are checked by a qualified clinical reviewer before use — extracted content is for reference only and not medical advice.Like a lobster shell, security has layers — review code before you run it.
latestvk972kapjs9mkqc007d3w83ym3583ex5s
License
MIT-0
Free to use, modify, and redistribute. No attribution required.