parse_traceback

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

No annotations are provided, so the description carries the full burden of behavioral disclosure. The description states it returns a structured output specifying exception type, message, frames, and the most probable cause frame (last user code frame, not library). This discloses the output format and heuristic behavior (cause detection), which is useful transparency for a parse function. It doesn't discuss edge cases (malformed input, empty traces) or side effects, but as a pure parsing tool that's less critical.

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 a single, front-loaded sentence that efficiently conveys the transformation behavior, output fields, and a key heuristic. It's compact with no filler. It could arguably benefit from an explicit statement of when to use it, but as a standalone technical description it's well-structured and concise.

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?

The tool has a single parameter with 0% schema description coverage, but the description clearly indicates the input is a raw Python traceback or pytest failure repr, which compensates well for the single unknown parameter. There is no output schema, and the description adequately enumerates the returned structure (exception type, message, frames, likely cause frame). Given the tool's modest complexity (one input, structured output), the description is reasonably complete, though it doesn't mention what happens with malformed or non-traceback input.

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?

While schema description coverage is 0%, there is only one parameter (traceback_text) whose semantics are well clarified by the description mentioning 'traceback Python brut (ou une repr d'echec pytest)'. The description explicitly defines what sort of text the parameter accepts, so the parameter meaning is effectively communicated even though not in the schema. This compensates for the 0% schema coverage.

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 uses a specific verb ('Transforme') with a concrete resource (raw Python traceback / pytest failure repr) and a clear output target (structured: exception type, message, localized frames, most likely cause frame). It clearly distinguishes from siblings like run_pytest, lint, and call_endpoint. The inclusion of the heuristic behavior (last user code frame rather than library frame) adds specificity beyond a generic 'parse' description.

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 implies its usage context (parsing tracebacks from failing pytest runs) but does not explicitly state when to use it versus alternatives or when not to use it. There are no sibling tools that obviously compete for this function among the listed names, so the lack of explicit exclusions is somewhat mitigated, but the description relies on inference rather than explicit guidance about its role in a debugging workflow.

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