The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to perform file discovery and system checks, including grep operations to detect TikZ diagrams and R code chunks within user-provided slide files.
[DYNAMIC_EXECUTION]: The skill executes a local script scripts/validate-findings.py using python3 via the Bash tool to validate JSON reports generated by sub-agents. This involves executing code provided within the skill's file structure.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted content from slide files (.tex, .qmd, .md) which is then passed to various sub-agents for analysis. This creates a vulnerability surface where malicious instructions embedded in the slides could potentially influence the sub-agents' behavior.
Ingestion points: User-provided filenames resolved in Step 1 and Step 2 of SKILL.md.
Boundary markers: The instructions do not specify the use of clear delimiters or instructions to ignore embedded commands when the agents process the slide content.
Capability inventory: The orchestrator and sub-agents have access to high-privilege tools including Read, Write, Bash, Agent, and Task.
Sanitization: There is no evidence of input sanitization or filtering of the slide content before it is processed by the agent lenses.

slide-excellence