The Agent Skills Directory

[DATA_EXFILTRATION]: The skill instructs the agent to inspect sensitive repository locations including "production environment files", "deployment configs", "CI / CD pipelines", and "save data". These files and directories are high-risk targets that frequently contain hardcoded secrets, API keys, or infrastructure credentials. Accessing these locations represents a data exposure risk, even though the skill does not include explicit network commands to transmit the data externally.
[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the repository being assessed, creating an attack surface for indirect prompt injection. Ingestion points: The agent reads potentially attacker-controlled content such as "release notes", "changelogs", "user-facing docs", "screenshots", "deployment configs", and "CI / CD pipelines" (SKILL.md). Boundary markers: There are no instructions providing delimiters or warnings to ignore potentially malicious embedded commands within these files. Capability inventory: The skill has the capability to write assessment results to the filesystem at docs/game-studio/project-state.md (SKILL.md). Sanitization: No sanitization, filtering, or validation steps are defined for the processed data before it is presented or written to the output file.

gsp-project-state-assessment