The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests and processes content from MVP.md and BRAND.md to automate the creation of backlogs and draft specifications. This functionality introduces a surface for indirect prompt injection, where an attacker could embed malicious instructions within the project documentation to influence the agent's behavior.
Ingestion points: The skill reads content from MVP.md and BRAND.md located in the project root or its Git superproject.
Boundary markers: There are no explicit instructions or delimiters defined to isolate the ingested data from the agent's primary instruction set.
Capability inventory: The skill has the capability to write files (writeFile) and execute local Node.js scripts via execFile.
Sanitization: The importar_mvp.mjs script performs a security check (containsSensitiveData) to identify and prevent the processing of hardcoded credentials like RSA keys and API tokens, though it does not specifically sanitize for prompt injection content.
[COMMAND_EXECUTION]: The skill performs various local operations by executing shell commands and scripts.
Evidence: SKILL.md instructs the agent to run node .agents/skills/specsfy-mvp-milestone-interviewer/scripts/importar_mvp.mjs and a similar capture script.
Evidence: The importar_mvp.mjs script uses child_process.execFile to execute Git commands (git rev-parse) and other Node.js scripts located in neighboring directories (iniciar_backlog.mjs, iniciar_spec.mjs).

specsfy-mvp-milestone-interviewer