The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill's primary function involves gathering intelligence from untrusted sources, including external websites (via the webfetch tool) and local codebase files, to synthesize a reconnaissance dossier. This process creates a surface for indirect prompt injection where malicious instructions embedded in those sources could influence the agent's behavior during the synthesis or subsequent workflow phases.\n
Ingestion points: External data fetched from the internet via the webfetch tool (analogues, prior art), local source code files, project manifests (e.g., package.json, requirements.txt), and ledger files in the docs/gangsta/ directory.\n
Boundary markers: The instructions lack specific requirements for using delimiters or 'ignore' instructions when interpolating gathered content into the dossier or agent context.\n
Capability inventory: The skill performs file system write operations to save dossiers and checkpoints, utilizes the webfetch network tool, and triggers the execution of subsequent automated workflow skills such as gangsta:the-grilling and gangsta:the-sit-down.\n
Sanitization: There are no instructions provided for sanitizing, filtering, or escaping the external content before it is processed or presented to the user.

reconnaissance