The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted external data, such as binary metadata, strings, and decompiled code, creating a potential surface for indirect prompt injection attacks where malicious instructions are hidden within the analyzed data.
Ingestion points: Data enters the context through analysis tools like get-strings, get-symbols, and get-decompilation as outlined in Phase 1 and Phase 2 of the workflow.
Boundary markers: The instructions lack specific boundary markers or "ignore" directives to prevent the agent from accidentally executing instructions found within the data it analyzes.
Capability inventory: The skill utilizes tools to modify the analysis environment, including rename-variables, set-decompilation-comment, set-bookmark, and checkin-program.
Sanitization: No sanitization or validation of the input data from the binary analysis tools is specified.
[NO_CODE]: The skill consists entirely of markdown instruction files and does not include any accompanying scripts, executables, or package dependency files.

ctf-pwn