Skip to content

Point agents at s1 feedback when a command fails, release 2.1.2 - #15

Merged
fatwang2 merged 1 commit into
masterfrom
claude/project-thread-uo0pv2
Oct 29, 2026
Merged

fatwang2 merged 1 commit into
masterfrom
claude/project-thread-uo0pv2

Conversation

@fatwang2

Copy link
Copy Markdown
Member

Requested by fatwang2 · project thread

Follows the Agentback onboarding guide (step 3b): wherever the CLI prints errors, also print feedbackHint on stderr, so agents learn they can report the problem with s1 feedback.

  • src/index.ts: under Error: ..., print If this looks like a problem with s1, report it: s1 feedback "<what happened>" (skipped when the failing command is feedback itself).
  • Update agentback to 0.3.0 (adds the web widget; agentback/cli and agentback/mcp are unchanged).
  • Version 2.1.2 + CHANGELOG. Release via the manual Release workflow after merge.
  • New test in tests/commands.test.mjs; npm test passes locally (31/31).

Generated by Claude Code

Print the Agentback SDK's feedbackHint under the error line, as the
Agentback onboarding guide recommends, and update agentback to 0.3.0.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01DuYf7tHNVauyjtcZzUzzgZ
@fatwang2 fatwang2 self-assigned this Oct 29, 2026
@fatwang2
fatwang2 merged commit 702a9ce into master Oct 29, 2026
2 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants