The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses uv run to execute a local Python script (scripts/sync_context.py). This script is used to process and summarize local configuration files, which is consistent with the skill's stated administrative purpose.
[INDIRECT_PROMPT_INJECTION]: The skill operates as a data aggregator, which creates a surface for potential indirect instruction injection if source files are compromised.
Ingestion points: The script reads user-defined content from identity.md, preferences.md, rules.md, and projects.md within the ~/.claude/.context/core/ directory.
Boundary markers: The generated file uses standard Markdown headers (e.g., ## Identity, ## Rules) to separate sections but does not employ specific escaping or "ignore instructions" delimiters for the content being aggregated.
Capability inventory: The skill performs file system read operations on personal context files and write operations to ~/.claude/rules/elle-core.md via the Python script.
Sanitization: There is no sanitization of the markdown content read from the source files; the content is concatenated and written directly to the target rules file.

sync-context