The Agent Skills Directory

[PROMPT_INJECTION]: The skill's core functionality involves reading and analyzing user-provided text or markdown files to generate titles, summaries, and formatting. This creates an indirect prompt injection surface where instructions hidden within the input files could manipulate the agent's behavior. \n
Ingestion points: Step 1 involves reading a user-specified file. \n
Boundary markers: No explicit delimiters or warnings are used when the agent processes the file content. \n
Capability inventory: The skill can execute shell commands (backups, typography scripts), read/write files, and perform complex content analysis. \n
Sanitization: The input data is processed without sanitization or instructions to ignore embedded commands. \n- [COMMAND_EXECUTION]: The skill uses shell execution for file operations and typography processing. \n
Evidence: SKILL.md contains Bash snippets using test to check paths and mv to create timestamped backups. \n
Evidence: scripts/autocorrect.ts uses execSync to run npx autocorrect-node on the processed file path. While the path is quoted, execution via a shell remains a risk if filenames contain metacharacters. \n- [EXTERNAL_DOWNLOADS]: The skill relies on runtime package retrieval via npx. \n
Evidence: SKILL.md instructs the agent to run the main script using npx -y bun. \n
Evidence: The typography script executes npx autocorrect-node to apply spacing fixes, which fetches the tool from the NPM registry at runtime.

baoyu-format-markdown