[instructions] Sync instruction files with release v0.82.2#43990
Conversation
Accuracy: document the `evals` frontmatter field (only undocumented top-level schema key); rescope `github-token`/`bots` under `on:` in syntax-agentic.md. Duplication: collapse the cost-aware triage bullet sets in the creator and updater prompts to reference the canonical High-Volume Triage and Escalation Pattern. Cleanup: fix orphaned report-incomplete paragraph in safe-outputs-runtime.md. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
This comment has been minimized.
|
Hey One flag from the automated checklist:
If you'd like to address this with an agent: Warning Firewall blocked 1 domainThe following domain was blocked by the firewall during workflow execution:
network:
allowed:
- defaults
- "patchdiff.githubusercontent.com"See Network Configuration for more information.
|
|
🎉 This pull request is included in a new release. Release: |
Summary
Syncs four instruction files under
.github/aw/with release v0.82.2. Changes are documentation-only — no Go code, compiled output, or workflow YAML is modified.Changed Files
.github/aw/create-agentic-workflow.md.github/aw/safe-outputs-runtime.md.github/aw/syntax-agentic.mdevals:field.github/aw/update-agentic-workflow.mdChange Details
create-agentic-workflow.md — §8 cost-aware triage
workflow-patterns.md#high-volume-triage-and-escalation-pattern.workflow-patterns.mdfrom the trailing "See also" list (now linked inline); keptsubagents.mdandtoken-optimization.md.safe-outputs-runtime.md — report-incomplete description
report-incompleteparagraph.syntax-agentic.md — field corrections and new evals: field
github-token:— description changed from "Default GitHub token for workflow" to "GitHub token override"; added note that it is not a top-level field and must be placed underon:,tools.github, orsafe-outputs.bots:— renamed toon.bots:to reflect correct frontmatter placement.evals:— new field (experimental). Documents BinEval binary (YES/NO) evaluation questions that run after safe-outputs and before the conclusion job. Supports shorthand form and extended form withquestions, optionalmodel, andruns-on.update-agentic-workflow.md — Cost-Oriented Update Checks section
workflow-patterns.md#high-volume-triage-and-escalation-pattern,subagents.md, andtoken-optimization.md.Agentic Analysis Notes
bots:toon.bots:andgithub-token:scope clarification are documentation corrections aligning docs with existing runtime behaviour.evals:is an experimental frontmatter field; its addition here is documentation only.workflow-patterns.mdis now referenced inline rather than via trailing "See also" links in two files.