fix: only load project context (CODEACTOR.md/CLAUDE.md) on first conv…

[iohub]

Summary by Sourcery

Cache project context loading per session and adjust when it is injected into the system prompt, refine TUI tool call rendering with bordered headers, and ignore .git changes in the codebase watcher while removing DevOps-Agent docs from the READMEs.

Enhancements:

• Cache project workspace context files in the Conductor agent so they are loaded only once per session and injected only on the first conversation turn.
• Update TUI tool call rendering to draw subtle borders around completed tool call headers while leaving running calls and bodies unbordered.
• Filter file watcher events in the code analysis service to ignore changes originating from the .git directory.

Documentation:

• Remove DevOps-Agent documentation sections from both English and Chinese READMEs.

[sourcery-ai]

Reviewer's Guide

This PR tweaks the TUI tool-call rendering to add thin borders around completed tool call headers, introduces reusable border helpers, caches project context so it is only loaded and injected into the system prompt once per conversation, and filters out .git directory events from the codebase watcher, while also removing the DevOps-Agent docs from the READMEs.

Sequence diagram for project context loading on first conversation

sequenceDiagram
    actor User
    participant ConductorAgent
    participant Memory as MemoryConversation
    participant FS as FileSystem
    participant Publisher
    participant LLM as LLMEngine

    User->>ConductorAgent: Run(ctx, input, mem)
    ConductorAgent->>Memory: GetMessages()
    Memory-->>ConductorAgent: messages (len == 0)
    alt first_conversation_turn
        ConductorAgent->>ConductorAgent: loadProjectContext()
        alt cachedProjectContext is nil
            ConductorAgent->>FS: read CODEACTOR.md / CLAUDE.md / AGENTS.md
            FS-->>ConductorAgent: contents
            ConductorAgent->>ConductorAgent: build Content and LoadedFiles
            ConductorAgent->>ConductorAgent: cache in cachedProjectContext
        else cachedProjectContext already set
            ConductorAgent-->>ConductorAgent: reuse cachedProjectContext
        end
        ConductorAgent->>Publisher: Publish(context_loaded, loadResult, Name())
        ConductorAgent->>ConductorAgent: prepend project context to systemPrompt
    else followup_conversation_turn
        ConductorAgent-->>ConductorAgent: skip loadProjectContext()
    end
    ConductorAgent->>LLM: send messages with systemPrompt
    LLM-->>ConductorAgent: response
    ConductorAgent-->>User: reply

Loading

File-Level Changes

Change	Details	Files
Add bordered headers for completed tool calls in the TUI and introduce reusable border helpers/styles.	Adjust RenderToolLine so early and running states remain borderless while successful tool calls get bordered headers. Ensure tools with skipped bodies still render within bordered headers. Wrap headers with new addToolCallBorders helper and keep bodies outside the border. Introduce ToolCallBorderTop/Bottom styles and a makeBorderLine helper for compact border lines.	internal/tui/render.go internal/tui/styles.go
Cache project context loading and only inject project context into the system prompt on the first conversation turn.	Add cachedProjectContext field to ConductorAgent to store the first load result. Update loadProjectContext to short-circuit on cached results and to cache newly loaded content. Gate project context injection in Run so it only happens when there are no prior conversation messages (first turn).	internal/agents/conductor.go
Ignore file change events coming from the .git directory in the codebase watcher.	Add filtering logic in the watcher to skip events whose paths include a .git component before queueing re-analysis.	codebase/src/http/handlers/mod.rs
Remove DevOps-Agent documentation sections from English and Chinese READMEs.	Delete DevOps-Agent section and its configuration/disable instructions from README.md. Delete DevOps-Agent section and its configuration/disable instructions from README_zh.md.	README.md README_zh.md

---

Tips and commands

Interacting with Sourcery

• Trigger a new review: Comment @sourcery-ai review on the pull request.
• Continue discussions: Reply directly to Sourcery's review comments.
• Generate a GitHub issue from a review comment: Ask Sourcery to create an
  issue from a review comment by replying to it. You can also reply to a
  review comment with @sourcery-ai issue to create an issue from it.
• Generate a pull request title: Write @sourcery-ai anywhere in the pull
  request title to generate a title at any time. You can also comment
  @sourcery-ai title on the pull request to (re-)generate the title at any time.
• Generate a pull request summary: Write @sourcery-ai summary anywhere in
  the pull request body to generate a PR summary at any time exactly where you
  want it. You can also comment @sourcery-ai summary on the pull request to
  (re-)generate the summary at any time.
• Generate reviewer's guide: Comment @sourcery-ai guide on the pull
  request to (re-)generate the reviewer's guide at any time.
• Resolve all Sourcery comments: Comment @sourcery-ai resolve on the
  pull request to resolve all Sourcery comments. Useful if you've already
  addressed all the comments and don't want to see them anymore.
• Dismiss all Sourcery reviews: Comment @sourcery-ai dismiss on the pull
  request to dismiss all existing Sourcery reviews. Especially useful if you
  want to start fresh with a new review - don't forget to comment
  @sourcery-ai review to trigger a new review!

Customizing Your Experience

Access your dashboard to:

• Enable or disable review features such as the Sourcery-generated pull request
  summary, the reviewer's guide, and others.
• Change the review language.
• Add, remove or edit custom review instructions.
• Adjust other review settings.

Getting Help

• Contact our support team for questions or feedback.
• Visit our documentation for detailed guides and information.
• Keep in touch with the Sourcery team by following us on X/Twitter, LinkedIn or GitHub.

[sourcery-ai]

Hey - I've left some high level feedback:

• In setup_watcher, the .git filter currently checks components for exact equality with .git; if you expect nested worktrees or alternative Git directories, consider making this filter more robust (e.g., checking for .git directories via metadata or a configurable ignore list) so you don’t inadvertently miss or over-filter events.
• The makeBorderLine helper is currently only invoked with the '─' character and its comment mentions half-block characters; either update the comment to match the actual usage or allow callers to pass a half-block rune so the intent and implementation stay aligned.

Prompt for AI Agents

Please address the comments from this code review:

## Overall Comments
- In `setup_watcher`, the `.git` filter currently checks components for exact equality with `.git`; if you expect nested worktrees or alternative Git directories, consider making this filter more robust (e.g., checking for `.git` directories via metadata or a configurable ignore list) so you don’t inadvertently miss or over-filter events.
- The `makeBorderLine` helper is currently only invoked with the `'─'` character and its comment mentions half-block characters; either update the comment to match the actual usage or allow callers to pass a half-block rune so the intent and implementation stay aligned.

---

Sourcery is free for open source - if you like our reviews please consider sharing them ✨

• X
• Mastodon
• LinkedIn
• Facebook

_{Help me be more useful! Please click 👍 or 👎 on each comment and I'll use the feedback to improve your reviews.}