The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to execute shell commands using eval on the output of a local script (~/.claude/skills/econstack/bin/econstack-slug). This is a risky pattern as it allows for the execution of dynamically generated code at runtime.
[EXTERNAL_DOWNLOADS]: The instructions recommend that the user clone and update a repository from an unverified third-party source (github.com/charlescoverdale/econstack-data.git) to populate a local parameter database. This source is not recognized as a trusted organization.
[COMMAND_EXECUTION]: The skill performs multiple executions of binaries located in a non-standard path (~/.claude/skills/econstack/bin/) during its initialization and completion phases for update checks and telemetry logging.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection due to its core functionality of processing external data.
Ingestion points: External data is ingested from the web using the WebSearch and WebFetch tools (SKILL.md).
Boundary markers: None identified; fetched content is interpolated directly into the report generation process.
Capability inventory: The skill has access to Bash for shell command execution, Write for file system modifications, and network tools (WebSearch, WebFetch).
Sanitization: There are no instructions for sanitizing or filtering untrusted content before it is processed by the agent.

market-research