Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

Annotations already declare read-only, local, idempotent, non-destructive, but the description adds the crucial operational facts: static and local, nothing fetched, run, or stored, one call, no auth. It also pre-empts the most dangerous misreading by defining that 'no-findings' means only that the listed patterns were absent from the submitted files, not that the repo is safe, and that 'unknown' means nothing was analyzed.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Is the description appropriately sized, front-loaded, and free of redundancy?

Front-loaded with the action and the verdict framing, then detection scope, then input limits, then return shape and the caveats — a sensible order. It is dense but long, and the Binance theft/recruitment-test backstory plus the near-duplicate restatement of the 40-file/28000-character caps already in the schema are the sentences that earn their place least.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a single-parameter tool with no output schema, the description supplies everything an agent needs: input shape and limits, which files to send, the four possible verdict values with their meanings, the findings record fields, the presence of coverage and note, and the explicit non-guarantee of 'no-findings'. The limitations are stated rather than hidden.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema coverage is 100% and the schema documents the path-to-text map, so the baseline is met. The description goes beyond it by explaining the caps in prose (at most 40 files, 28000 characters total) and, more usefully, by naming the high-value paths to include — package.json, .vscode/tasks.json, .claude/settings.json, .mcp.json, .devcontainer/devcontainer.json, .yarnrc.yml, .npmrc, setup.py, plus any suspicious script — which is selection guidance the schema cannot express.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Does the description clearly state what the tool does and how it differs from similar tools?

States a specific verb and resource: pass a stranger's repository files in and get a verdict on run-on-open and install-time lures. It enumerates the exact pattern classes it detects (npm install hooks, VS Code run-on-open tasks, .claude/.cursor/.mcp.json agent hooks, devcontainer initializeCommand, yarnPath, git/URL deps, decode-and-eval payloads, setup.py fetches), which clearly separates it from siblings like check-package and check-workflow.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Does the description explain when to use this tool, when not to, or what alternatives exist?

Explicitly frames the when: before opening or running a repo someone else sent you, with concrete triggers (recruiter coding test, downloaded zip, PoC). It also names the alternative paths for a different scope — the command line scanner or the DugganUSA Checks GitHub Action for a whole directory — and states the safety prerequisite of reading code on a credential-free machine.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.