The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from the local repository, including specification files (SPEC.md), ticket descriptions (tickets/*.md), commit logs, and source code, to generate the handoff report. Malicious instructions embedded in these external sources could potentially influence the agent's reasoning or the content of the generated report.
Ingestion points: SPEC.md, tickets/NN-*.md, git log, and various source files.
Boundary markers: Absent. There are no explicit delimiters or instructions telling the agent to ignore embedded commands within the processed repository data.
Capability inventory: File system read access (via git and file reading tools) and file system write access (saving reports to the .scratch/ directory).
Sanitization: Absent. The skill does not perform escaping, validation, or filtering of the content read from the repository before it is processed as natural language.

frontend-handoff