Pharmaclaw Chemistry Query
v2.0.1Chemistry agent skill for PubChem API queries (compound info/properties, structures/SMILES/images, synthesis routes/references) + RDKit cheminformatics (SMIL...
⭐ 3· 764·4 current·4 all-time
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description match the included scripts: PubChem/ChEMBL/PubMed queries, RDKit processing, visualization, retrosynthesis, and reaction templates. Declared external APIs and system dependency (optional Java/OPSIN) are used by the code, and there are no unrelated credentials or surprising binaries requested.
Instruction Scope
SKILL.md instructs use for compound lookups, RDKit analysis, and retrosynthesis, which the scripts implement. It also mentions chaining outputs to a 'pharma-pharmacology-agent' (a usage suggestion rather than an autonomy requirement). The scripts run subprocesses and access only local skill files, call public APIs (PubChem/ChEMBL/NCBI), and write visualizations under a local viz directory; they do not attempt to read or exfiltrate unrelated system files or require secrets.
Install Mechanism
There is no formal install spec; code runs as-is. One runtime behavior worth noting: opsin_name_to_smiles.py will auto-download an OPSIN JAR from a GitHub release on first use. The download is verified with a pinned SHA-256 checksum in the script, reducing risk. No other remote arbitrary downloads or unverified installers are present.
Credentials
The skill declares no required environment variables or credentials and the code does not attempt to access hidden tokens. External APIs used are public (no key required). The lack of requested secrets is proportionate to its stated functionality.
Persistence & Privilege
always is false and the skill does not request permanent system-wide presence or modify other skills' configurations. It writes images/reports into its own viz/scripts directories and runs local subprocesses — expected for this type of tool.
Assessment
This skill appears coherent and implements what it claims: PubChem/ChEMBL/PubMed queries plus RDKit-based analysis, drawing, and retrosynthesis. Before installing or running it, consider: (1) it can generate multi-step synthesis routes and contains named reaction templates and experimental conditions — if your organization restricts assistance for lab synthesis, review the content carefully; (2) the OPSIN JAR is downloaded at runtime from a GitHub release but the script verifies a pinned SHA-256 (good practice); (3) the skill runs multiple Python scripts as subprocesses and writes visualization files into a viz directory under the skill — check file-write locations and sanitize any sensitive environment where you run it; and (4) no secrets are required. If you need higher assurance, review the included files (particularly templates.json and rdkit_reaction.py) for content you consider sensitive and run in an isolated environment (container or sandbox).Like a lobster shell, security has layers — review code before you run it.
latestvk973nnfz50c2q78gh1pnaxmhnn82q1tz
License
MIT-0
Free to use, modify, and redistribute. No attribution required.