analyze-elf

HIGH W007: Insecure credential handling detected in skill instructions.

• Insecure credential handling detected (high risk: 0.90). The skill explicitly extracts raw strings and JS source from binaries (string extraction, /treesitter) and emits JSON/walkthrough outputs, so any embedded secrets (API keys, tokens, passwords) in the binary would likely be output verbatim.

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (medium risk: 0.30). In analyze-elf’s features command, the skill reads free text from the user-provided binary by running --help (help_text = run_cmd([binary, "--help"], ...)) and then parses those command description strings (_parse_help_commands(help_text)), so an outsider can embed prompt-injection text in the binary’s help output that the pipeline consumes at runtime.