The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill defines a workflow where the agent ingests untrusted user data (geometry dimensions, material properties, support/load descriptions) which is intended to be used as parameters for Abaqus-related tasks involving shell commands and file writing.
Ingestion points: SKILL.md defines "Required Information" including dimensions, material names, support types, and load magnitudes provided by the user.
Boundary markers: The skill does not define specific delimiters or instructions for the agent to treat these inputs as data rather than potential instructions.
Capability inventory: The skill frontmatter allows access to Bash(abaqus:*), Write, Edit, Read, Glob, and Grep tools.
Sanitization: The instructions lack any guidance on validating, escaping, or filtering user-provided parameters before they are interpolated into scripts or commands.
[METADATA_POISONING]: The skill contains conflicting information regarding its source and authorship.
Evidence: UPSTREAM.md attributes the source to a repository by majiayu000, while metadata.json attributes the author and source to JaimeCernuda. Furthermore, the environment context identifies the author as cai-aa. These inconsistencies, while potentially stemming from repository forks, obscure the authoritative provenance of the skill instructions.

abaqus-static-analysis