Add guardrails for large command output and potentially sensitive Git output so maintainers can use the server more safely in AI-assisted workflows.
Acceptance criteria:
- Define output-size limits for tools that can return large diffs, logs, or file lists.
- Warn when output may contain secrets, tokens, private remotes, or local filesystem details.
- Prefer truncation with clear continuation guidance over unbounded output.
- Include tests for truncation and warning behavior.
Add guardrails for large command output and potentially sensitive Git output so maintainers can use the server more safely in AI-assisted workflows.
Acceptance criteria: