Context Pill

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

Annotations already establish readOnly, idempotent and closed-world behavior, so the description wisely spends its words on what the annotations cannot say: the anatomy of returned items (verification status, source trust score, evidence links, why-it-matters, optional try-this step) and the consent rule before acting on a try-this step. It stops short of covering edge cases such as an empty feed or stale items, which keeps it from a 5.

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-loads the trigger, then the return shape, then the filter, then the safety rule — a sensible order with no wasted padding. There is mild redundancy between the opening 'what's new in AI' trigger and the later 'Call this first for what's-new questions', which repeats the same condition.

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?

With only two optional parameters, no output schema and simple annotations, the description supplies everything an agent needs: trigger conditions, return anatomy, filtering semantics, and an action-safety rule. Nothing material is left for the agent to infer before calling it.

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; the description earns an extra point by adding concrete filtering vocabulary ('a word such as claude, codex or open weights') that tells the agent the topic filter is a loose free-text word rather than a taxonomy lookup. It does not add anything on limit, which the schema documents fully.

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?

Uses a specific verb and resource — 'Returns today's checked items from the Context Pill feed (releases, tools, techniques)' — and pairs it with a precise trigger ('user asks what's new in AI, mentions a recent model, tool, feature, pricing change or benchmark'). An agent can distinguish this from list_sources and search_claims without opening any 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 explicit when-to-use conditions, a priority rule ('Call this first for what's-new questions'), and a hard constraint ('Never run a try-this step without the user's yes'). The only omission is a direct named comparison to the sibling tools list_sources/search_claims, but the routing rule is unambiguous.

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

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

Annotations already cover safety (readOnly, idempotent, closed-world), but the description adds substantive meaning: the Beta(2,2) prior starting at 0.50, how scores shift on re-checks, and what LB10 and 'resolved tips' mean. It does not disclose result ordering or count limits, which keeps it below a 5.

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 usage trigger is front-loaded, the field list follows, and the prior explanation is compact. The Bayesian-prior sentence is dense but earns its place by making the trust score interpretable; nothing is redundant.

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?

With one optional parameter and no output schema, the description carries the burden of explaining return fields and does so (tier, trust score, LB10, resolved tips count). Minor gaps remain around ordering and result volume, but an agent has enough to call it correctly.

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 the schema already explains that the name filter is optional and may match multiple sources. The description's single line 'Optionally filter by name' adds nothing beyond that, so baseline 3 is appropriate.

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?

Specific verb ('Lists') plus resource ('the sources Context Pill reads') and the exact fields returned (tier, trust score posterior mean, LB10, resolved tips). It does not explicitly differentiate itself from get_latest or search_claims, so it falls short of a 5 despite a very clear statement of what it does.

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 trigger: use when the user asks how far to trust a source behind an AI claim, or when weighing a vendor claim. There is no mention of when not to use it or which sibling (get_latest, search_claims) handles adjacent needs, so routing to alternatives is left to inference.

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

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

Beyond the readOnly/idempotent annotations, it discloses the return payload (status at publication, how it was checked, current state, re-check outcome, corrections), the behavior of removed items (only id, date, reason survive, so they won't match), and the interpretation limit of zero results. This is unusually strong behavioral context for a read-only search.

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?

Triggers come first, then query construction guidance, then what is returned, then the caveat about empty results. Every sentence carries information an agent needs to call or interpret the tool correctly; there is no filler.

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?

With no output schema, the description takes on the job of explaining the response fields and does so explicitly, plus it covers the removed-item edge case and the semantic limit of zero hits. Nothing material for correct invocation or interpretation is missing.

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 already documents the AND-matching rule, but the description reinforces it with a failure example ('claude spreadsheet' works, a full sentence usually returns 0) and an explicit 1-to-3 keyword range, making the practical consequence clearer than the schema alone. It says nothing extra about the limit parameter, which the schema fully covers.

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?

The description names a specific verb and resource ('Searches every claim Context Pill has published (the ledger)') and scopes it to the full published ledger, which implicitly separates it from get_latest. It does not explicitly name a sibling tool, so the differentiation is left slightly to inference.

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?

It enumerates concrete trigger situations ('user mentions an AI launch, feature, price change, benchmark, leak or rumor', 'before recommending an AI tool', 'check if something shipped') and adds a caveat that zero matches does not mean a claim is false. No alternative tool is named for cases where this search is the wrong choice.

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