The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands, specifically python3 $PAPERCUT_LIB/tools/catalog.py. The use of environment variables in command paths can lead to unintended execution if the environment is not properly controlled.
[DATA_EXFILTRATION]: The skill accesses the file path ~/.config/papercut-explainer/config.env to resolve environment variables. Access to the ~/.config directory and files containing environment configurations is considered sensitive and could lead to data exposure.
[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes untrusted data from narration/script.json, which is then used to generate storyboard content and trigger build tools.
Ingestion points: The agent reads the contents of narration/script.json as the primary source for the storyboard.
Boundary markers: There are no explicit delimiters or instructions provided to the agent to ignore potentially malicious instructions embedded within the JSON script.
Capability inventory: The skill has the capability to execute shell commands (python3, papercut-build) and perform file system writes.
Sanitization: There is no evidence of validation, escaping, or sanitization performed on the input data before it is used in the storyboard generation process.
[DYNAMIC_EXECUTION]: The skill executes a Python script from a path ($PAPERCUT_LIB/tools/catalog.py) that is dynamically resolved at runtime from environment variables or configuration files. This pattern of dynamic loading from computed paths poses a risk if the configuration source is compromised.

papercut-storyboard