Code Trace Tree: Sync Agent Traces into Your IDE

Where this fits

You're an agent that just finished refactoring a function across three files. You've updated the source, but the user's IDE still shows the old trace points — the line numbers are stale, and the tree structure no longer matches the code. The user asks: "Can you update my trace tree to reflect the new flow?" Without a way to read and write the plugin's data, you'd have to hand them a list of manual steps. The code-trace-tree skill closes that gap: it lets an agent operate the Code Trace Tree plugin's hybrid storage directly, then trigger a reload in the IDE.

Why agents benefit

• Precise trace point management – Add, move, or delete line/file/directory trace points with exact locators (--file --line --content) while the script computes occurrence automatically.
• Nested workflow generation – Use create_tree.py to build a full workflow tree in one go; it ensures existing nodes and only adds new ones, avoiding redundant searches.
• Source-aware rebinding – After editing source on disk, run trace_tree rebind so LINE locations stay aligned with the new code, preventing silent drift.
• IDE integration – The skill knows how to ask the IDE to reload plugin data and can navigate the tree to a specific trace point, making the update visible immediately.
• Safe by default – It only edits traces when the user explicitly asks, and it understands that writing under <OS Config Dir>/code-trace-tree/ is expected, not a workspace violation.

Practical scenarios

Refactoring a critical path

The user asks you to refactor a payment validation flow. You update the source files, then use trace_tree rebind to realign existing line traces, and finally move a file-level trace point to the new module. The IDE tree updates instantly, and the user can click through the new flow.

Generating a fresh trace tree from a spec

Given a high-level description of a workflow, you use create_tree.py to generate a nested tree of trace points that mirrors the logical steps. You don't need to search for existing nodes — the script handles that. The user sees a structured map of the process in the IDE.

Removing obsolete traces after a cleanup

After deleting a legacy module, you remove all trace points that referenced it. You use trace_tree delete with a directory locator, then trigger a reload so the IDE doesn't show dangling entries.

Add it to your agent workflow

Install the skill via npx:

npx skills add saidake/code-trace-tree-jetbrains --skill code-trace-tree

Then, in your agent prompt, you can say: "Add a line trace at src/utils.js:42 with content const result = await process()" and the skill will handle the rest. After any source modification, remember to run trace_tree rebind to keep locations fresh.

Compare before adopting

Before committing, check the skill's quality signals: the repository has 59 stars and an active author (saidake). Look at how often it's updated and whether the underlying plugin supports your IDE version. Alternatives might be manual file editing or a custom script, but those lack the rebinding logic and IDE reload integration. Also consider whether your workflow involves frequent trace tree updates — if not, the overhead might not be worth it.

Why it is worth tracking

This skill is a niche but practical example of agent-IDE collaboration. Its small but focused feature set, clear scripts, and explicit safety guardrails make it a solid candidate for teams that build agents for code navigation and refactoring. The community momentum is modest but real, and the skill demonstrates a pattern — bidirectional sync between agent and IDE plugin — that will become more common. Evaluate it if you're building agents that need to keep developer tooling in sync with code changes. For a broader look at similar skills, see OpenAgentSkill's listing.