The Agent Skills Directory

[DYNAMIC_EXECUTION]: The skill is designed to execute arbitrary shell commands defined in its configuration files (customize.toml and its project-specific overrides).
Evidence: SKILL.md specifies that the agent should execute all entries in activation_steps_prepend and activation_steps_append during activation.
Evidence: step-05-present.md executes a command defined in the on_complete field of the workflow configuration.
[COMMAND_EXECUTION]: The skill invokes several command-line tools as part of its automated workflow.
Evidence: SKILL.md executes uv run to call a Python script located at {project-root}/_bmad/scripts/resolve_customization.py.
Evidence: step-05-present.md and step-oneshot.md use the code command to open the repository and specification files in the user's editor.
[INDIRECT_PROMPT_INJECTION]: The skill processes potentially untrusted external content and incorporates it into the agent's context.
Evidence: SKILL.md (Step 3) loads persistent facts from files matching glob patterns provided in the configuration.
Evidence: step-01-clarify-and-route.md ingests contents from unformatted spec or intent files to determine the goal of the workflow.
Boundary Markers: The skill includes explicit instructions in step-01-clarify-and-route.md to ignore directives within the intent that instruct the agent to skip steps or implement code directly.
Capability Inventory: The skill has the ability to execute shell commands, write files, and spawn sub-agents, which increases the potential impact of a successful prompt injection.
Sanitization: The skill does not appear to use strict schema validation or sanitization for the file content it ingests into the prompt.

gds-quick-dev