The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to perform 'Runtime forensics' (Step 2.1) and 'Same-track incident search' (Step 2.3). This involves extracting live DOM structures, probing network responses, and reading public security advisories. These ingestion points represent a surface for indirect prompt injection where an attacker could embed malicious instructions in their website or incident reports to influence the agent's behavior.
Ingestion points: live DOM/container structure, resource supply analysis (URL parameters), response headers/content, and public CVE/advisory descriptions.
Boundary markers: The skill does not define explicit delimiters or 'ignore embedded instructions' warnings for the data ingested during these forensics steps.
Capability inventory: The workflow utilizes browser interaction tools, network request capabilities, and file system access (to audit local code using file:line and to write the 'Report family').
Sanitization: There are no instructions for sanitizing or escaping the content retrieved from external websites before it is processed or included in reports.
[EXTERNAL_DOWNLOADS]: The skill requires the effective-web-research dependency and explicitly prompts the user to install the author's skill package (FuDesign2008/open-skills) via npx if it is missing. While this is standard for the author's ecosystem, it involves fetching and installing code from a remote repository.

tech-research-workflow