The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes multiple scripts and binaries located in the user's home directory (~/.claude/skills/econstack/bin/), including econstack-update-check, econstack-learnings-read, econstack-learnings-log, and econstack-greenbook.
[EXTERNAL_DOWNLOADS]: The skill references and instructs the user to fetch data and code from a non-trusted repository at https://github.com/charlescoverdale/econstack-data.git via git clone and git pull.
[REMOTE_CODE_EXECUTION]: The skill performs dynamic code execution by using eval "$(~/.claude/skills/econstack/bin/econstack-slug)". This executes the output of a script directly as shell commands, which is a high-risk pattern if the script's behavior is influenced by external or untrusted data.
[PROMPT_INJECTION]: The skill processes untrusted external markdown files through the --from argument.
Ingestion points: Reads project, options, costs, and benefits data from external .md files in Step 1.
Boundary markers: No explicit boundary markers or 'ignore' instructions are used when parsing these files.
Capability inventory: The skill has extensive capabilities including Bash execution, Write access to the file system, and the ability to invoke other Skill tools.
Sanitization: There is no evidence of validation or sanitization of the content imported from the markdown files before it is used in calculations or shell command construction.

cost-benefit