One comment per agent run #38

New Issue

2026-07-04T10:36:21+02:00

ffaerber commented

2026-07-04 10:36:21 +02:00

Currently if a agents run it creates multiple comments in the PR. Better is when the agents only creates one comment per run. This is also including the tool calls & reasoning

🚀 1

gitea-actions bot commented

2026-07-04 10:36:40 +02:00

🤖 @pm

The current behavior of agents creating multiple comments per run (including reasoning and tool calls) creates noise in the PR thread. We need to modify the agent execution loop to buffer all output and post a single consolidated comment at the end of the run.

Since this involves modifying the core execution logic and how the agent interacts with the Gitea API to post comments, this is a complex task requiring an understanding of the agent's lifecycle.

Plan:

Identify the logic responsible for posting comments during an agent's run (likely in the workflow or the agent wrapper).
Implement a buffering mechanism to collect reasoning, tool calls, and final responses.
Update the logic to post the aggregated content as a single comment once the agent finishes its task or reaches a terminal state.
@qa will verify that only one comment is created per run and that all relevant information is preserved.

I propose @senior to implement this change.

@ffaerber ready to start building? reply yes to proceed.

🤖 **@pm** The current behavior of agents creating multiple comments per run (including reasoning and tool calls) creates noise in the PR thread. We need to modify the agent execution loop to buffer all output and post a single consolidated comment at the end of the run. Since this involves modifying the core execution logic and how the agent interacts with the Gitea API to post comments, this is a complex task requiring an understanding of the agent's lifecycle. **Plan:** 1. Identify the logic responsible for posting comments during an agent's run (likely in the workflow or the agent wrapper). 2. Implement a buffering mechanism to collect reasoning, tool calls, and final responses. 3. Update the logic to post the aggregated content as a single comment once the agent finishes its task or reaches a terminal state. 4. @qa will verify that only one comment is created per run and that all relevant information is preserved. I propose @senior to implement this change. @ffaerber ready to start building? reply yes to proceed.

ffaerber commented

2026-07-04 10:39:55 +02:00

@pm yes

🚀 1

gitea-actions bot commented

2026-07-04 10:40:10 +02:00

🤖 @pm

The plan has been approved. I am now handing this task over to @senior to implement the output buffering and consolidate agent comments into a single post per run.

🤖 **@pm** The plan has been approved. I am now handing this task over to @senior to implement the output buffering and consolidate agent comments into a single post per run.

ffaerber commented

2026-07-04 10:40:10 +02:00

@senior please proceed with issue #38 per the plan above (delegated by pm).

🚀 1

gitea-actions bot commented

2026-07-04 10:40:17 +02:00

🔨 @senior is on it — building on branch ai/issue-38. I'll open a PR when it's ready.

🔨 **@senior** is on it — building on branch [`ai/issue-38`](https://git.ffaerber.duckdns.org/ffaerber/agents/src/branch/ai/issue-38). I'll open a PR when it's ready.

gitea-actions bot referenced this issue from a commit

2026-07-04 10:42:58 +02:00

feat(publish): one comment per agent run (issue #38)

gitea-actions bot referenced a pull request that will close this issue

2026-07-04 10:43:22 +02:00

@senior: One comment per agent run #39

gitea-actions bot referenced this issue

2026-07-04 10:43:23 +02:00

@senior: One comment per agent run #39

ffaerber closed this issue

2026-07-04 10:55:40 +02:00

Sign in to join this conversation.