Add daily-storify workflow for episode-based daily narrative analysis

[Copilot]

This adds a new daily Agentic Workflow, Storify, that converts the last 24h of github/gh-aw workflow activity into a blog-style narrative. It focuses on cross-workflow episodes, emergent feedback loops, and the human interventions that changed outcomes.

• Workflow contract (new file)

  • Introduces .github/workflows/daily-storify.md with daily + manual triggers, read-only analysis permissions, and tracker-id: daily-storify.
  • Defines a narrative output focused on episode chains, loop dynamics, and intervention impact rather than raw run summaries.

• 24h data acquisition + evidence model

  • Reuses shared 24h log prefetch via shared/aw-logs-24h-fetch.md.
  • Requires correlation across pre-fetched logs, agentic-workflows MCP (logs/audit), and GitHub activity (issues/PRs/reviews/discussions) to ground the story in traceable evidence.

• Durable notes via cache-memory

  • Adds explicit cache-memory persistence under:
    • /tmp/gh-aw/cache-memory/storify/state.json
    • /tmp/gh-aw/cache-memory/storify/episodes.jsonl
    • /tmp/gh-aw/cache-memory/storify/loops.jsonl
  • Preserves loop taxonomy and carry-forward questions across runs to improve continuity and reduce re-discovery.

• Output behavior

  • Publishes one discussion using daily-audit safe outputs ([storify] prefix, 2d expiry).
  • Enforces explicit noop when no credible cross-workflow loop can be established and report_incomplete only for real blockers.

• Compiled artifact

  • Adds generated .github/workflows/daily-storify.lock.yml to keep markdown source and runtime workflow in sync.

imports:
  - shared/aw-logs-24h-fetch.md
  - uses: shared/daily-audit-base.md
    with:
      title-prefix: "[storify] "
      expires: 2d

tools:
  agentic-workflows:
  cache-memory: true

[Copilot]

Pull request overview

Adds a daily Storify workflow that narrates cross-workflow episodes, feedback loops, and human interventions from the prior 24 hours.

Changes:

• Adds scheduled/manual narrative analysis with durable cache memory.
• Publishes findings through constrained discussion safe outputs.
• Includes the generated runtime workflow.

Show a summary per file

File	Description
.github/workflows/daily-storify.md	Defines Storify’s tools, analysis process, memory, and output contract.
.github/workflows/daily-storify.lock.yml	Compiled GitHub Actions workflow.

Review details

Suppressed comments (2)

.github/workflows/daily-storify.md:155

• The agent-facing safe-output tool is named create_discussion (as shown in the generated <safe-output-tools> list), not create-discussion. Using the frontmatter-style hyphenated name here can make the agent invoke a nonexistent tool instead of publishing the story.

- Use `create-discussion` when you have a valid story.

.github/workflows/daily-storify.md:157

• report_incomplete is not exposed to this agent: the generated manifest and prompt list only create_discussion, missing_data, missing_tool, and noop. A tooling blocker therefore cannot follow this terminal path. Expose report_incomplete through the compiler/tool list, or change this contract to the available missing-tool/data signals and regenerate the lock file.

- Use `report_incomplete` only for real tooling/infrastructure blockers.

• Files reviewed: 1/2 changed files
• Comments generated: 2
• Review effort level: Balanced

[github-actions]

🎉 This pull request is included in a new release.

Release: v0.87.5