From software problem to a clear next step.
Choose a topic and what you need to do next. Add a search phrase to find a specific problem.
Browse 20 topics
Explore the playbooks
1500 guides · Page 15 of 60
-
Triage an Unclear Agent Brief Before Work Starts
Turn an ambiguous agent request into a bounded brief by identifying the decision, inputs, constraints, output, owner, and missing context.
-
Prioritize Work From an Unclear Agent Brief
Choose whether to clarify, narrow, investigate, or defer an agent task using decision value, evidence readiness, urgency, and scope uncertainty.
-
Investigate Why an Agent Brief Is Ambiguous
Trace an unclear agent request from source wording to task setup and output expectations to identify the exact ambiguity and missing evidence.
-
Verify an Agent Brief Is Actionable and Complete
Check that an agent brief names one decision, bounded scope, evidence, constraints, deliverable, review owner, and a clear stopping condition.
-
Learn From Repeated Unclear Agent Briefs
Capture recurring ambiguity patterns, brief examples, ownership, and review triggers so agent tasks become easier to start and evaluate.
-
Triage an Agent Workflow With Missing Evidence
Identify the decision, required observations, available artifacts, and evidence gaps before treating an agent result as ready for human review.
-
Prioritize Evidence Collection for an Agent Task
Rank evidence gaps by decision consequence, claim uncertainty, collection cost, freshness, and the chance that one artifact resolves several open questions.
-
Investigate Missing Evidence in an Agent Recommendation
Trace each recommendation claim to its source and comparison, exposing omitted artifacts, unsupported inferences, and unresolved evidence boundaries.
-
Verify Evidence Sufficiency for an Agent Handoff
Check that each material claim has traceable, timely evidence, clear uncertainty, and enough comparison for a human to review the agent result.
-
Learn From Repeated Missing Evidence in Agent Work
Turn evidence gaps into claim ledgers, source expectations, review fixtures, and recurrence signals that improve agent handoffs without overstating certainty.
-
Triage Contradictory Agent Recommendations
Compare recommendation claims, inputs, evidence, criteria, and timing first so conflicting agent outputs can be scoped before choosing a winner.
-
Prioritize Conflicting Agent Recommendations
Rank recommendation conflicts by decision consequence, disagreement scope, evidence quality, recurrence, and the cost of delaying a human resolution.
-
Investigate Why Agent Recommendations Conflict
Reconstruct conflicting agent runs with matched inputs, evidence, criteria, and timing to identify the first divergence and its decision effect.
-
Verify Consistency of Agent Recommendations
Check that comparable agent runs use the same question, evidence, criteria, and boundaries, and that legitimate alternatives remain clearly labeled.
-
Learn From Contradictory Agent Recommendations
Capture disagreement patterns, matched cases, decision ownership, and recurrence review so conflicting agent outputs become explainable workflow signals.
-
Triage Repeated Investigation in an Agent Workflow
Compare prior briefs, questions, evidence, and stopping conditions to determine whether repeated investigation adds coverage or only repeats the same work.
-
Prioritize Repeated Investigation Versus New Evidence
Choose whether to merge, stop, clarify, or continue an agent investigation using decision value, new evidence, changed context, and remaining uncertainty.
-
Investigate Why an Agent Repeats the Same Work
Trace prior task context, evidence indexes, handoffs, and stop rules to explain repeated investigation and identify the missing workflow boundary.
-
Verify an Agent Investigation Adds New Evidence
Check that a repeated investigation has a changed question or input, produces a named evidence delta, and stops when its decision boundary is resolved.
-
Learn From Repeated Agent Investigations
Capture overlap fingerprints, handoff fields, stop rules, and recurrence review so agent investigations add evidence instead of cycling through the same path.
-
Triage Stale Context in an Agent Workflow
Compare context timestamps, revisions, scope, and current observations to decide whether an agent result is still relevant or needs a bounded refresh.
-
Prioritize Refresh Work for Stale Agent Context
Rank context refreshes by decision consequence, change likelihood, artifact age, freshness requirement, and the amount of work a bounded update can replace.
-
Investigate Why an Agent Used Stale Context
Trace context collection, task setup, revisions, and source selection to explain how outdated material reached an agent result and affected its claims.
-
Verify Agent Context Freshness Before Review
Check agent inputs against the freshness cutoff, make revisions visible, label historical context, and confirm changed facts affect conclusions appropriately.
-
Learn From Repeated Stale Agent Context
Capture context manifests, freshness rules, revision examples, and recurrence review so agent results stay aligned with the system they describe.