harness_execute

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

Beyond the annotations (readOnlyHint=false, destructiveHint=true, idempotentHint=false), the description discloses concrete behavior: killing/restoring feature flags (what can be destroyed), server-side blocking wait until terminal status, URL-based auto-extraction of identifiers, and parallel HQL batch execution. The parameter-level text reinforces this with a risk/confirm model ('required when the action's risk is medium_write or above'), confirmation semantics ('Does NOT override an explicit decline'), and timeout behavior ('returns execution_timed_out=true with the last observed status'). 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?

Four sentences, all high-signal: purpose with action examples, the URL shortcut, the wait-until-terminal pattern, and the HQL special case. The broadest statement is front-loaded, followed by the two highest-value invocation tips. No filler, no restatement of schema content — appropriately compact for a tool with 16 parameters and a 60+ value resource_type enum.

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 highly complex tool (16 params, large enums, nested objects, output schema present), the definition is complete: return values are covered by the output schema, the safety profile by annotations, parameter semantics by 100% schema coverage, and the non-obvious invocation patterns (wait, HQL batch, URL extraction) by the description. The pointer to harness_describe for action-specific fields closes the otherwise unbounded action/resource-type matrix.

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 each parameter description is already rich (defaults, min/max, sibling references, 'Ignored for other actions'), so the baseline is 3. The main description adds value by encoding cross-parameter dependencies that prevent mis-invocation: wait only applies to pipeline run/retry, queries is only valid with resource_type='hql_query' AND action='validate' or 'run', and URL auto-extracts identifiers across the resource_type/org/project/id fields. These rules exist in the schema but are scattered; the description synthesizes them into callable patterns.

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 action verb and resource — 'Execute an action on a Harness resource' — and grounds it with concrete examples: run/retry/interrupt pipelines, kill/restore FME feature flags, test connectors, sync GitOps apps, run chaos experiments. These action verbs clearly differentiate it from CRUD siblings (harness_create/update/delete) and read siblings (harness_get/list/search), so an agent can route to it without opening schemas.

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?

Gives explicit when-to-use conditions in the main prose: 'Pass wait: true for pipeline run/retry to block until the execution reaches a terminal status — single tool call instead of an LLM polling loop' and 'For HQL batch operations pass queries with resource_type="hql_query" and action="validate" or "run"'. Prerequisite routing to siblings is embedded in parameter descriptions ('Check runtime_input_template first via harness_get', 'Call harness_describe for available fields', 'List available: harness_list'). However, it never explicitly excludes CRUD operations — the boundary with harness_create/update/delete is implied by the action examples but never stated as a when-not-to-use.

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