ingestSource

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

With no annotations, the description carries the full burden and does most of it well: it discloses that the tool writes data, that it is idempotent via content-hash comparison, that project scope is refused and why (plaintext vs. sensitive external content), the default caps, and that dryRun acts as a consent gate. The gap is job lifecycle — it returns a jobUid but doesn't say whether the run is synchronous or must be polled via listJobs/waitForIdle, nor what failure mid-ingest leaves behind.

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?

Front-loaded with the core action and its incremental nature, then moves to inputs, then to guarantees and return value — a sensible order. Dense but nearly every clause carries information; the stacked parentheticals in the parameter sentence make it slightly harder to scan than it needs to be.

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 six-parameter mutation with an output schema, this covers prerequisite state, connector-specific setup, scope safety, defaults, and idempotency — enough to invoke correctly. What's missing is how the returned jobUid is meant to be used (job tracking via the sibling job tools), which matters given the ingest is potentially long-running and bounded by maxItems/maxExtractionCalls.

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 baseline is 3; the description adds real meaning by grouping parameters by connector type (repoPath applies to file-based connectors, omitted for granola) and by characterizing dryRun as a consent gate and maxItems as a backfill bound. It also flags the one hard constraint that matters most — scope only accepts personal/team and project is refused — before the agent reads the enum. Minor redundancy with the schema's own descriptions keeps it from a 5.

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 ("Ingest the delta from a previously connected source, resuming from its saved watermark") and immediately frames the operation as incremental rather than full-sync. The phrase "previously connected source" distinguishes it from connectSource, and the connector list makes the operating domain concrete without opening a schema.

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 clear selection context: use this for a cheap incremental pull against an existing connection, with an explicit exclusion ("Refuses project scope") and per-connector setup guidance (granola needs a keychain key set via 'maz source add granola', file-based ones need repoPath). It never names what to call when no source is connected yet (presumably connectSource) or how to follow up on the returned jobUid, so the alternative-routing guidance is implied rather than explicit.

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