The Agent Skills Directory

[SAFE]: The skill utilizes command-line tools (gjalla) and specific MCP resources that belong to the skill author's legitimate ecosystem. These are used for standard development tasks such as scaffolding specifications and tracking progress.
[SAFE]: The instructions implement strong security posture by requiring explicit human approval gates before the agent can move from the planning phase to the implementation phase ("Do not begin implementing without explicit approval").
[INDIRECT_PROMPT_INJECTION]: The skill identifies a surface for indirect prompt injection as it processes external data from a master specification and project source code.
Ingestion points: External architecture diagrams, capabilities, data flows, and source code files are fetched into the agent's context using the gjalla CLI or MCP.
Boundary markers: The skill does not explicitly define delimiters for this external content, though it treats the data as reference material for a plan presented to the user.
Capability inventory: The skill has the capability to execute specific shell commands via the gjalla CLI (spec new, loop start, ledger mark).
Sanitization: There is no explicit sanitization of the input data described in the skill; however, the output of the planning process is subject to manual human review and approval before execution.

gjalla-spec