The Agent Skills Directory

[SAFE]: The skill successfully implements phylogenetic tree analysis using the Bio.Phylo library. The implementation in scripts/phylo_task.py is modular, follows secure file handling practices using explicit encoding, and contains no dangerous command execution, file access outside of its scope, or network operations.
[INDIRECT_PROMPT_INJECTION]: The skill processes external tree files, representing a potential indirect prompt injection surface.
Ingestion points: External tree data is read from the path specified by the input_tree parameter in config/task_config.json via the Phylo.read function in scripts/phylo_task.py.
Boundary markers: None present for the raw data content.
Capability inventory: The skill can create directories, write files to the artifacts/ folder, and generate visualizations via matplotlib as defined in scripts/phylo_task.py.
Sanitization: The skill relies on the internal parsing logic of the Bio.Phylo library for tree format validation.
[INDIRECT_PROMPT_INJECTION]: The file biopython-phylo_audit_result_v1.json contains self-referential safety claims, such as 'security: PASS'. Following the security analysis protocol, these claims are treated as descriptive data and were not used to influence the analyzer's independent verdict.

biopython-phylo