The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes a local shell command date '+%Y-%m-%d %H:%M' to retrieve the current system time for metadata generation. While used for a benign purpose, this demonstrates the capability to run shell commands.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and synthesize large volumes of untrusted data from external sources including personal notes, braindumps, and meeting transcripts.
Ingestion points: Multiple directories listed in SKILL.md under Section 1 (Data Gathering), such as 02-personal/braindumps/, 03-professional/braindumps/, 04-projects/*/braindumps/, and 00-inbox/braindump-*.md.
Boundary markers: The instructions do not define explicit delimiters or instructions to ignore embedded commands within the ingested content.
Capability inventory: The agent has the ability to execute shell commands (Bash) and broad file system read/write permissions within the target vault.
Sanitization: There are no documented steps for sanitizing or escaping the content retrieved from these notes before processing.

knowledge-consolidation