loom_context_pack

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

Annotations already declare readOnlyHint/idempotentHint/destructiveHint=false, so the safety profile is covered; the description adds real behavior beyond that — the bundle is prompt-sized (bounded), what each section contains, and that include_runs yields a manifest with goal, outcome, tool calls and files. It stops short of describing truncation behavior when token_budget is exceeded or ordering of sections, which is the remaining gap.

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 purpose before the conditional include_runs detail and the sibling routing, so an agent can stop reading early. Dense but no filler; the parenthetical run-manifest enumeration is the only slightly heavy clause, and it earns its place.

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 15-parameter, zero-required aggregation tool with no output schema, the definition covers what comes back and one key toggle but leaves the mode enum semantics (startup vs task vs debug vs handoff) undefined, which is the single most consequential knob. Adequate to call it correctly in the default case, incomplete for deliberate configuration.

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 only 13% against 15 parameters, so the description must carry weight: it does explain include_runs concretely and implicitly maps the section list to only_sections, and hints that query/situation drive selection. However mode (startup/task/debug/handoff), read_scope, namespace, run_id, top_k, runs_limit, token_budget, include_chunks and the include_* flags are left entirely to the schema, so coverage is partial rather than compensating.

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 resource — a prompt-sized context bundle containing profile, relevant memories, matching procedures and optional docs/conversations — and names the two siblings it is not (loom_search for a single fact, loom_activate for activation only). An agent can distinguish it from every retrieval sibling 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 an explicit timing rule ('once near the start of a substantial task') and two explicit exclusions with the alternative to use instead ('For one fact, prefer loom_search; for activation only, loom_activate'). It also sequences a follow-up action ('shortlist there before opening one with memory.evidence or memory.digests'), so when-to-use and when-not-to-use are both covered.

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