The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute genmedia CLI commands for model discovery, schema inspection, and executing media generation jobs.
[EXTERNAL_DOWNLOADS]: Generated files are downloaded from remote service providers to the local path ./outputs/characters/ using the --download flag in genmedia commands.
[DATA_EXFILTRATION]: User-provided reference images are uploaded to external endpoints (e.g., fal-ai/, openai/) to facilitate identity-preserving generation tasks.
[PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface where user-provided character attributes and prompt variables are interpolated into shell command strings.
Ingestion points: Character identity anchors and shot variables collected from user input as described in SKILL.md and related reference files.
Boundary markers: The prompt structure uses descriptive headers like CHARACTER ANCHOR but does not implement security-hardened delimiters to separate untrusted data from instructions.
Capability inventory: The agent has the capability to execute shell commands (genmedia) and perform local file writes via the --download flag.
Sanitization: There is no evidence of sanitization or escaping of shell metacharacters for the prompt strings passed to the CLI tool.

character-design