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

Far exceeds the annotations (destructiveHint, idempotentHint, readOnlyHint=false) by disclosing exactly what gets destroyed (free-text history/checkpoints/task/context of referencing sessions, outcomes, links, jobs, linked evidence), what survives (shared entities still referenced), idempotency semantics (repeat execution returns a minimal receipt), a content-fingerprint safeguard against replay repopulation, and quota consumption.

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 purpose and alternative, then proceeds through prerequisites, dry-run workflow, and scope exclusions. It is dense and long, but nearly every sentence carries a distinct operational fact (irreversibility, destruction scope, fingerprint safeguard, quota, out-of-scope stores), so waste is minimal.

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 high-stakes destructive operation, the description covers permissions, irreversibility, exact destruction scope, surviving data, execution workflow, idempotency behavior, and explicit out-of-scope exclusions. With an output schema present, return-value detail is optional, and the description is more than complete.

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 100%, so dry_run, memory_id, and confirmation are already documented in the schema, including the exact confirmation format and the warning not to substitute a repo name. The description restates this workflow (dry_run default true, execute with the reviewed ID) but adds little syntax or meaning beyond the schema, so the baseline of 3 applies.

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 ('Preview or permanently erase one private memory and its explicit derived patterns from the live database') and explicitly distinguishes itself from the sibling engram_retract_memory. An agent can tell exactly what this does and how it differs from alternatives without opening the 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?

Explicit when-to-use ('Inspect it first'), how-to-execute ('Execute only the reviewed ID using dry_run=false and confirmation="delete:<memory_id>"'), the alternative it is not ('Unlike engram_retract_memory, execution cannot be undone'), and prerequisites ('Requires a human workspace OWNER API key; OAuth agents and member keys cannot erase'). Exclusions are also stated ('Does not delete independent memories or unrelated sessions').

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