Playbooks

Playbooks · Agent context

Clear old tool results after processing

Keep recent results while removing re-fetchable payloads from long-running context.

context-editing tool-result-clearing

Playbooks · Agent workflows

Look up customer/order details and draft a reply with read-only tools

Use explicit client-side tools for synthetic customer/order lookups and draft a reply without cancelling or changing anything.

customer-support read-only

Playbooks · Agent workflows

Turn multi-source IOC lookups into a threat-intelligence brief

Query simulated intelligence sources for synthetic IP/hash/domain indicators, correlate findings, and produce a sourced draft.

threat-intelligence mock-data

Playbooks · Agent evaluation

Generate privacy-safe synthetic test cases for a prompt template

Generate diverse inputs within template-variable constraints, then compare prompt versions on them.

synthetic-data prompt-evaluation

Playbooks · Agent workflows

Query independent tools in parallel within one analysis turn

Let Claude query independent data sources together and synthesize results, avoiding serial waits.

parallel-tool-calls latency

Playbooks · Agent workflows

Ground an architecture map in an unfamiliar repository

Compare source code with stale architecture docs in a disposable repo, then add evidence to the live session.

Playbooks · Agent workflows

Fix a failing test fixture and verify independently

Iterate on failing tests in a locked-down fixture, then rerun assertions outside the agent session.

Playbooks · Agent workflows

Carry an issue through CI and review to a verified PR

Recover from both a CI failure and reviewer feedback instead of assuming a linear workflow.

Playbooks · Agent workflows

Isolate each user’s SaaS credential in a separate vault

Create a vault for a fictional user, attach a bearer credential, verify agent access, and clean up resources.

Playbooks · Agent workflows

Drive a long-running task through webhooks and resume human approval

Return quickly from HTTP requests, receive run events by webhook, pause for review, and resume the session.

Playbooks · Agent workflows

Turn a CSV into a chart-backed analysis report

Upload a sales CSV and produce a single HTML report with interactive charts and numeric evidence.

data-analysis reporting

Playbooks · Agent workflows

Deliver CSV analysis in a Slack thread with follow-ups

Connect an @mention and CSV attachment to an analysis session, stream progress in-thread, and return the report there.

slack data-analysis

Playbooks · Agent workflows

Cap session spend and resume safely

Bound the cost of an unattended research session, inspect state at the cap, and resume only when authorized.

spend-control

Playbooks · Agent workflows

Grade a research brief with an independent rubric

Have a writer produce evidence, then let an independent grader check coverage and source quality.

artifact-review research

Playbooks · Agent workflows

Canary-test and roll back a prompt version

Compare prompt versions on labeled examples and pin production back to the old version after regression.

prompt-evaluation rollback

Playbooks · Agent workflows

Plan research, then parallelize fact verification

Have a coordinator decompose a question, bounded workers verify authoritative sources, and synthesize one report.

research multi-agent

Playbooks · Agent workflows

Coordinate specialists into a coherent proposal

Assign domain questions to specialist agents and synthesize a consistent decision proposal.

multi-agent planning

Playbooks · Agent workflows

Create an accessible narration from a chart-heavy PDF deck

Pass a presentation as a multimodal PDF and produce a structured narration that explains charts.

multimodal document-analysis

Playbooks · Agent workflows

Turn mixed handwritten forms into reviewable JSON

Extract form fields and transcription while surfacing uncertainty for human confirmation.

document-extraction vision

Playbooks · Agent workflows

Submit and reconcile offline message batches

Submit many independent prompts asynchronously, then reconcile successes and failures by `custom_id`.

batch-processing

Playbooks · Agent workflows

Report team usage, model mix, and costs

Aggregate token usage and USD costs by period and workspace through the Admin API into an auditable CSV.

billing analytics

Playbooks · Agent workflows

Compare tool designs against a task set

Run agent tool calls on a fixed task set and measure success, latency, and tool feedback.

tool-evaluation testing

Playbooks · Agent workflows

Reproduce a search-agent benchmark and score accuracy

Run a search agent on a fixed question set and quantify results with reference answers and a defined grader.

evaluation search

Playbooks · Agent workflows

Compare context and tool-cost optimization on benchmark cases

Compare lean context, selective retrieval, subagents, and batching under the same quality threshold.

cost-optimization evaluation

Playbooks · Agent workflows

Pin an inference region for regulated workloads

Set `inference_geo` when creating an agent, constrained by the workspace allowlist, and verify session behavior.

deployment compliance