harvest_did

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

Even though annotations already declare readOnlyHint=true and openWorldHint=false, the description adds substantial behavioral context beyond them: explicit statistical assumptions (parallel trends per 2x2 cell, no anticipation at reference -1, independence across units, cross-horizon covariance ignored), pre-conditions, and two named failure modes with remediation paths. No contradiction with annotations.

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 long but well-structured with clear labeled sections (Assumptions, Pre-conditions, Failure modes, Alternatives) and the core purpose is front-loaded in the first sentence. Each section earns its place for a complex 15-parameter estimator, though it is denser than strictly necessary.

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 complex staggered-DID estimator with 15 parameters, an output schema, and annotations, this description covers purpose, assumptions, data requirements, failure modes, and alternatives — everything an agent needs to decide whether to call it and how to interpret its behavior. The output schema covers return values, so no description burden there. Only minor loss is the truncated 'Alternatives: sp.callawa...' sentence.

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%, so the baseline is 3, but the description adds meaning beyond the schema by clarifying the treat/cohort distinction (dynamic vs static harvesting), explaining never_value as a never-treated marker, and tying reference horizon -1 to the 'clean' anticipation assumption. Minor gap: it references weighting options ('equal', 'n_treated') and 'cohort-weighted' aggregation that do not appear as explicit schema parameters, which could leave an agent unable to act on that guidance.

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: 'Harvest every valid 2x2 DID comparison from a staggered panel and aggregate them via precision-weighted / simple / cohort-weighted averages.' It differentiates from the large DID sibling family by declaring itself agnostic to cohort structure and positioning it as a robustness check 'against CS / SA / BJS', so an agent can distinguish it from callaway_santanna, sun_abraham, staggered_sa, and did_2x2.

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?

Provides explicit use-case guidance ('useful for robustness comparisons against CS / SA / BJS'), names alternatives, and gives operational pre-conditions (long-format panel, treat-or-cohort columns, multiple cohorts with overlapping periods) plus failure modes with concrete remedies (add never-treated units, restrict horizons, switch weighting). This is actionable when-to-use and when-not-to-use guidance.

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