Repository navigation
Conversation
Coding agents (and people) can report what blocked or misled them:
beam feedback -m "deploy ignored --port" --kind bug --blocked
The report goes to the account API of the selected context
({auth_url}/cli/feedback, as `beam setup` does for checkout) with the
workspace token and the CLI's caller and agent-session headers. Failures
carry codes agents can branch on: INPUT_REQUIRED, INVALID_ARGS (naming the
flag to fix), NOT_AUTHENTICATED, RATE_LIMITED, UNAVAILABLE and
FEEDBACK_DISABLED. BEAM_NO_FEEDBACK=1 turns the command off.
Co-authored-by: Cursor <cursoragent@cursor.com>
beam setup agent appends a Feedback section to the use-beam skill once beam-client depends on a beta9 with SDKSettings.skill_sections (beam-cloud/beta9#2092). Older beta9 releases ignore it. Co-authored-by: Cursor <cursoragent@cursor.com>
mernit
force-pushed
the
eli/agent-feedback
branch
from
October 31, 2026 16:52
c93a100 to
8561001
Compare
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Coding agents, and people, can now tell us when Beam blocks or misleads them:
The command posts to the selected context's account API at
{auth_url}/cli/feedback. That'shttps://api.beam.cloud/v2/oauth/cli/feedbackin production, and staging or local with--context. It sends the workspace token, asbeam setupdoes for credit checkout. internal-api (beam-cloud/internal-api#343) turns each report into a PostHog event. Nothing here or in beta9 stores feedback.-m/--messageis required, and is prompted for in a terminal. The rest are--kind(bug, error_message, missing_capability, docs, performance, praise, other),--blocked,--command,--error,--details,--suggest,--model,--format jsonand--context.x-beta9-caller,x-beta9-agent-session), so reports from one agent run group together. Once beam-cloud/beta9#2087 ships, the caller also names the agent, e.g.cli/0.1.293 (claude-code).{"sent": true}with--format json.BETA9_JSON=1:INPUT_REQUIREDINVALID_ARGS, which names the flag to fix, e.g.--message: At most 300 characters; put the rest in details.NOT_AUTHENTICATEDandUNAVAILABLEBEAM_NO_FEEDBACK=1makes the command refuse withFEEDBACK_DISABLED.feedback.SKILL_SECTIONtells agents when to report (Beam's fault, not the user's), what to include, and to carry on afterwards.beam setup agentappends it to theuse-beamskill throughSDKSettings.skill_sections(beam-cloud/beta9#2092).Release order
Release this after internal-api#343 is deployed; until then the endpoint returns 404. There's no version bump here.
Testing
tests/cli/test_feedback.pyhas 18 tests, passing with beta9 0.1.291 (the locked version). They cover:beam feedback --helpthrough the real CLIWith #2092's beta9 on
PYTHONPATH, theuse-beamskill installed with these settings ends with the rendered Feedback section.I also ran it end to end against internal-api#343 running locally on port 8008, using
--context local:feedback.submittedevent with every field, plus the caller and session headers.--format jsonprinted{"sent": true}.--message.