The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill provides instructions for the agent to install the GEF (GDB Enhanced Features) tool using a piped shell execution pattern: bash -c "$(curl -fsSL https://gef.blah.cat/sh)" in the file tools-advanced-2.md.
[EXTERNAL_DOWNLOADS]: The skill references multiple external tools and repositories for installation, including:
https://github.com/zrax/pycdc in SKILL.md.
https://github.com/Lil-House/Pyarmor-Static-Unpack-1shot in languages.md.
https://github.com/ohos-decompiler/abc-decompiler in languages.md.
https://github.com/pwndbg/pwndbg in SKILL.md.
https://github.com/mandiant/GoReSym in languages-compiled.md.
[COMMAND_EXECUTION]: The skill extensively uses shell commands for tool installation and binary analysis, including pip install, apt install, git clone, chmod +x, and various debuggers/analyzers like gdb, radare2, and frida.
[INDIRECT_PROMPT_INJECTION]: The skill has a significant attack surface for indirect prompt injection as it is designed to ingest and process untrusted data from CTF challenges.
Ingestion points: The agent is instructed to read and analyze arbitrary binary files (binary), APK files (app.apk), and WASM modules (checker.wasm) as referenced in SKILL.md, tools.md, and platforms.md.
Boundary markers: Absent. The instructions do not specify the use of delimiters or warnings when processing the contents of analyzed files.
Capability inventory: The skill allows access to Bash, Read, Write, Edit, WebFetch, and WebSearch tools as defined in the YAML frontmatter of SKILL.md.
Sanitization: Absent. There are no instructions to sanitize or validate the content extracted from target binaries before it is processed by the agent.

ctf-reverse