cos_graph_impact

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

Annotations already declare the tool as read-only, idempotent, and non-destructive. The description adds meaningful behavioral context: BFS-based traversal (depth and visit_limit), edge filtering by confidence_min, and the risk-tier grouping output. It also explains the truncation signal (meta.walk_truncated) that indicates when to increase limits. This goes beyond annotation information without contradicting it.

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?

The description is well-structured: a one-sentence purpose that front-loads the core function, followed by a compact Args block. All parameter explanations are precise and useful, with no filler or redundancy. The format makes it easy to scan and locate key details like uid schemes and defaults.

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?

Given the tool's complexity (5 parameters, graph traversal, risk tiers), the description is comprehensive. It covers every parameter with examples and edge cases, references a sibling tool for uid discovery, explains default behavior (confidence_min 0.3, visit_limit 500), and describes the risk-tier output. The presence of an output schema means return-format details are not needed from the description.

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?

The schema provides only parameter names and defaults (0% coverage), but the description fully documents each parameter in the Args block. uid includes a detailed scheme with examples and auto-resolution behavior; direction explains downstream vs upstream semantics; depth is defined as a BFS hop limit; confidence_min is tied to edge scores; visit_limit is clarified as a node-visit cap with a tuning hint. This far exceeds schema-level information.

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?

The description opens with a specific action ('Group affected nodes by risk tier') and names the concrete outcome categories (will_break / should_review / context). This clearly distinguishes it from sibling graph tools like cos_graph_query or cos_graph_trace, which focus on traversal or search rather than impact classification.

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?

The description explicitly advises running cos_graph_query first when unsure about the uid, naming a direct alternative. It also explains how to choose direction ('downstream (callers — break if uid changes)' vs 'upstream (deps uid calls/imports)') and provides a practical hint to raise visit_limit when meta.walk_truncated is true. This is actionable guidance for selecting and tuning the tool.

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