The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill facilitates the automated auditing of local source code files to identify deprecated Apollo API usage patterns, which establishes a surface for indirect prompt injection.\n
Ingestion points: Local source code files (e.g., .ts and .js files in the src directory) that are read and scanned by the audit scripts.\n
Boundary markers: Absent. There are no delimiters or specific instructions provided to the agent to treat the audited code as untrusted data or to ignore embedded instructions.\n
Capability inventory: The skill is configured with access to the Read, Grep, and Bash(curl:*) tools, which define the potential impact if an injection were to occur.\n
Sanitization: Absent. The scripts output matching code lines directly to the agent's context without filtering or encoding.\n- [DYNAMIC_CONTEXT_INJECTION]: The SKILL.md file includes a shell command that is executed dynamically when the skill is loaded by the agent.\n
Evidence: !npm list axios 2>/dev/null | head -5\n
Description: This command checks for the presence and version of the axios package in the local environment. This is a benign diagnostic command used for environment discovery and does not perform sensitive operations.\n- [COMMAND_EXECUTION]: The auditing scripts provided in the skill use execSync to construct and execute shell commands.\n
Evidence: execSync(grep -rl "${ep.pattern}" ${srcDir} ...) in api-audit.ts.\n
Description: The use of string interpolation to build shell commands using the srcDir parameter presents a command injection risk if the parameter is influenced by untrusted input.

apollo-upgrade-migration