Systematic diary exploration: discover tags, entry distribution, coverage gaps, agent mistakes, and compile recipes
86
90%
Does it follow best practices?
Impact
81%
1.06xAverage score across 5 eval scenarios
Advisory
Suggest reviewing before use
A team onboarding a new AI agent to their codebase needs the agent to quickly understand the state of the project's diary system. They want a single comprehensive report covering everything: what types of entries exist, what tagging conventions are in use, what incidents have occurred, how commit patterns look, what the diary doesn't cover, and what context pack recipes would be useful.
The team also needs the report to be resumable — if the agent's context window fills up mid-analysis, another agent should be able to pick up where the first left off. The report itself should be stored as a diary entry.
Create the following files:
exploration-report.ts — A TypeScript module that exports:
generateExplorationReport(data: ExplorationData): string function that produces a formatted markdown reportExplorationData type covering all sectionsfindRelationOpportunities(data: ExplorationData): RelationCandidate[] function that identifies cross-entry connections worth creatingreport-template.md — A template showing the exact structure of the exploration report with placeholder sections.
recovery-protocol.md — A document explaining the recovery procedure: how to detect if a prior exploration exists, which phases were completed, and how to resume from the next incomplete phase.