The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes current conversation history to extract facts, preferences, and rules, which are then persisted in core configuration files. This creates a surface for indirect prompt injection where malicious input could alter the agent's behavior.
Ingestion points: Reviewing the conversation history for user identity, preferences, and behavior rules (SKILL.md).
Boundary markers: No specific boundary markers or instructions to ignore embedded commands are mentioned when writing to destination files like rules.md or identity.md.
Capability inventory: The skill has the ability to write to files in ~/.claude/.context/core/ and execute local scripts via uv run (SKILL.md).
Sanitization: The instructions do not specify any sanitization or validation logic for the content extracted from the conversation before it is written to the filesystem.
[COMMAND_EXECUTION]: The skill uses shell commands to interact with the filesystem and execute internal scripts.
The skill executes ls ~/.claude/.context/ to verify the environment (SKILL.md).
The skill executes uv run python ${CLAUDE_PLUGIN_ROOT}/skills/sync-context/scripts/sync_context.py to synchronize context files (SKILL.md).

update-context