The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill's installation instructions direct the agent to fetch the skill file and repository from github.com/shuvonsec. This is a third-party source not recognized as a trusted vendor, posing a supply-chain risk if the repository is compromised.
[PROMPT_INJECTION]: The methodology contains strong behavioral overrides (e.g., 'STOP. Do not write. Do not explore further. Move on.') designed to force the agent to ignore any findings it deems 'theoretical.' While intended to improve report quality, these are technically overrides of the agent's default reasoning protocols.
[INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of untrusted external data by instructing the agent to use crawling and resolution tools like katana, waybackurls, and httpx.
Ingestion points: Data enters the context via the output of web crawlers and HTTP probing tools in SKILL.md workflows.
Boundary markers: The skill does not define explicit delimiters or instructions to ignore embedded commands within the data being processed.
Capability inventory: The skill has broad capabilities to execute shell commands, run Python/Go scripts, and interact with the local file system.
Sanitization: There is no evidence of sanitization or filtering of the external content before it is processed by the agent.
[DYNAMIC_EXECUTION]: The skill provides a significant number of shell command templates and instructions for the agent to execute third-party binaries (e.g., ffuf, nuclei, sqlmap) and dynamically install packages (pip3 install, go install). This creates a broad surface for command execution.
[METADATA_POISONING]: The skill's frontmatter description includes localized Chinese trigger words ('中文触发词'), which may be used to influence the agent's behavior or visibility in specific locales.
[REMOTE_CODE_EXECUTION]: The skill references book.hacktricks.xyz, a well-known security research wiki. Although flagged by automated scanners, this is a standard industry resource and is considered a false positive in the context of an offensive security skill.

bb-local-toolkit