rec_stats

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

Annotations only establish read-only/non-destructive, so the description adds real behavioral context: results are maintained live and incrementally in the normal case, but are recomputed from the recorded file when statistics were skipped under load, for earlier sessions, or when exact=true. That silently doubles as a cost/latency hint. It does not state how expensive the recompute is or whether an active session is required, keeping it short of a 5.

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?

Two dense sentences with no filler, front-loaded with the returned statistics before the execution-mode caveat. The parenthetical 'Live (incremental) while complete' is slightly compressed and takes a second read, but overall it earns its length.

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?

An output schema exists, so the description need not enumerate return values, yet the extra detail on run timelines and edge timestamps is still useful orientation. Combined with the live-vs-recompute semantics, an agent can call it correctly; the only residual gap is the four undocumented parameters.

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 description coverage is 0%, so the description bears the full parameter burden. It usefully disambiguates the most confusing parameter (exact=true forces file-based recomputation) and hints at the runs parameters via 'value runs (state timeline with dwell times)', but id, symbols, max_runs, and include_runs are never explained, leaving meaningful gaps.

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 names a specific verb+resource (per-symbol statistics over the whole recording) and then enumerates exactly which statistics are produced (count, first/last, min/max/mean, change count, BOOL edges, value runs with dwell times). That enumeration distinguishes it from adjacent recording tools like rec_pulses, rec_latency, or rec_read, which return raw samples rather than aggregates.

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?

It explains the operational context that decides the mode of computation (live incremental vs file recompute after skipped ingest under load, for earlier sessions, or with exact=true), which implicitly tells the agent when exact=true is worth setting. However, it never states when to prefer this tool over recording siblings (rec_read, rec_pulses, rec_snapshot, rec_compare) or any explicit 'use X instead' exclusion.

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