Skip to content

v2: Add tiered retry protocol, auth refresh, and escalation to recover-from-error #36

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
nam20485 wants to merge 1 commit into
base: main
Choose a base branch
from
Open

v2: Add tiered retry protocol, auth refresh, and escalation to recover-from-error #36

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
9 changes: 9 additions & 0 deletions ...on_modules/ai-workflow-assignments/orchestrate-dynamic-workflow-input-syntax.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -88,6 +88,15 @@ For flexible command wrappers:
- $assignment_inputs = `$ARGUMENTS[1]` # e.g., { workflow: "implement-epic" }
```

<!-- v2: Added error recovery, validation standards, and env considerations (2025-10-29) -->
### Error Recovery & Validation Standards
- On any failure: Follow tiered retry in recover-from-error.md (targeted fix → scratch retry → escalate after 3 attempts).

@gemini-code-assist gemini-code-assist Bot Oct 30, 2025

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

To improve navigation and make it easier for readers to find the referenced document, consider making recover-from-error.md a clickable link.

Suggested change
- On any failure: Follow tiered retry in recover-from-error.md (targeted fix → scratch retry → escalate after 3 attempts).
- On any failure: Follow tiered retry in [recover-from-error.md](./recover-from-error.md) (targeted fix → scratch retry → escalate after 3 attempts).

Copilot AI Oct 30, 2025

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The reference to 'recover-from-error.md' is not a complete URL or relative path. For consistency with the repository's linking patterns and to enable proper navigation, this should be a full link to the canonical remote repository following the pattern used elsewhere in the codebase (e.g., https://raw.githubusercontent.com/nam20485/agent-instructions/main/ai_instruction_modules/[subdirectory]/recover-from-error.md).

Suggested change
- On any failure: Follow tiered retry in recover-from-error.md (targeted fix → scratch retry → escalate after 3 attempts).
- On any failure: Follow tiered retry in [recover-from-error.md](https://raw.githubusercontent.com/nam20485/agent-instructions/main/ai_instruction_modules/recover-from-error.md) (targeted fix → scratch retry → escalate after 3 attempts).

Copilot uses AI. Check for mistakes.
- Validation: Always use a dedicated qa-test-engineer agent for post-assignment-complete events. Input: Execution outputs + real state query (e.g., GitHub API). Output: Independent PASS/FAIL with evidence (e.g., "Repo query confirms issue #X exists").
- Prompt Standard: Include explicit data interpolation (e.g., $full_epic_json) and validation check before delegation.
### Environment Considerations
- For write-heavy workflows (e.g., GitHub issues): Ensure gh CLI auth (gh auth login) before run. If in simulated env, auto-escalate to manual as per recover-from-error.

@gemini-code-assist gemini-code-assist Bot Oct 30, 2025

Copy link
Copy Markdown
Contributor

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

medium

The reference to recover-from-error for auto-escalation in a simulated environment could be confusing, as the recover-from-error.md document doesn't explicitly mention this specific scenario. This could lead to misinterpretation by the agent. To improve clarity, I suggest making the instruction more direct and self-contained.

Suggested change
- For write-heavy workflows (e.g., GitHub issues): Ensure gh CLI auth (gh auth login) before run. If in simulated env, auto-escalate to manual as per recover-from-error.
- For write-heavy workflows (e.g., GitHub issues): Ensure gh CLI auth (gh auth login) before run. If in a simulated environment where writes are not possible, escalate for manual intervention.

Comment on lines +93 to +97

Copilot AI Oct 30, 2025

Copy link

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Similar to the previous reference, 'recover-from-error' should be a complete link to the canonical remote repository. Incomplete references make it difficult for agents to resolve the correct file location as required by the Single Source of Truth Policy.

Suggested change
- On any failure: Follow tiered retry in recover-from-error.md (targeted fix → scratch retry → escalate after 3 attempts).
- Validation: Always use a dedicated qa-test-engineer agent for post-assignment-complete events. Input: Execution outputs + real state query (e.g., GitHub API). Output: Independent PASS/FAIL with evidence (e.g., "Repo query confirms issue #X exists").
- Prompt Standard: Include explicit data interpolation (e.g., $full_epic_json) and validation check before delegation.
### Environment Considerations
- For write-heavy workflows (e.g., GitHub issues): Ensure gh CLI auth (gh auth login) before run. If in simulated env, auto-escalate to manual as per recover-from-error.
- On any failure: Follow tiered retry in [recover-from-error.md](https://raw.githubusercontent.com/nam20485/agent-instructions/main/ai_instruction_modules/ai-workflow-assignments/recover-from-error.md) (targeted fix → scratch retry → escalate after 3 attempts).
- Validation: Always use a dedicated qa-test-engineer agent for post-assignment-complete events. Input: Execution outputs + real state query (e.g., GitHub API). Output: Independent PASS/FAIL with evidence (e.g., "Repo query confirms issue #X exists").
- Prompt Standard: Include explicit data interpolation (e.g., $full_epic_json) and validation check before delegation.
### Environment Considerations
- For write-heavy workflows (e.g., GitHub issues): Ensure gh CLI auth (gh auth login) before run. If in simulated env, auto-escalate to manual as per [recover-from-error.md](https://raw.githubusercontent.com/nam20485/agent-instructions/main/ai_instruction_modules/ai-workflow-assignments/recover-from-error.md).

Copilot uses AI. Check for mistakes.
- Test: Run with --dry-run flag to validate without writes.

## Prompt Guidance (Embed in any chat prompt)
Use this pre-execution preamble in any prompt that invokes this orchestrator to ensure consistent, acceptance-gated runs across tools and contexts.

Expand Down