What does this skill do, and when should you use it?
This skill finds and reads timestamped logs from the latest Code OSS or Agents app development run. It covers main-process, renderer, extension-host, agent-host, MCP, terminal, network, and related logs. It locates the newest run under the default or explicitly supplied user-data directory and selects files according to the investigation. It is suited to developers who already have local development-build logs and need to diagnose startup, extension, agent, MCP, terminal, or network problems.
Identifies whether the request concerns Code OSS or the Agents app; locates the logs directory under the default or custom user-data directory; runs ls -lt to find the newest timestamped run folder; lists its contents; reads files such as main.log, window1/renderer.log, agenthost.log, extension-host logs, MCP logs, terminal logs, or network logs according to the issue; uses tail for recent entries, rg for error or probe searches, and find for non-empty logs. For temporary console.log probes, it also documents how to enable development-console forwarding temporarily, reproduce the issue, and restore the default disabled state.
- A Code OSS developer investigating startup failures or crashes needs main.log and renderer logs.
- An extension developer investigating activation or runtime failures needs extension-host and per-extension logs.
- An Agents app user diagnosing Copilot, agent-session, or agent-host communication issues needs agenthost.log and related output logs.
- A developer troubleshooting MCP, terminal, network authentication, or settings-sync problems needs the corresponding process logs.
- A source contributor using temporary probes needs console output forwarded into normal logs and the temporary changes removed afterward.
How do you install this skill?
- Assume logs may contain sensitive data and redact content before sharing, pasting, or uploading it.
- Enabling temporary forwarding and probes changes source code and increases log exposure; confirm scope and ensure the default state is restored even after interruption.
- The procedure assumes standard user-data directories and Unix-style commands; Windows, restricted permissions, empty logs, and custom launch configurations may require additional handling.
- Shell / CLI
- Local filesystem
The repository contains 61 skills; this skill is located at .github/skills/code-oss-logs/SKILL.md. The README does not provide a standalone installation command or installer, so the exact installation method is undocumented. If the client supports Agent Skills directories, place the skill directory containing this file according to the client's directory convention; that convention is not specified in the source.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/microsoft/vscode.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/.github/skills/code-oss-logs" ~/.claude/skills/
rm -rf "$tmp"Generated from the source repository and skill path; it copies only this skill's folder. If the author's install steps above differ, follow those first. To scope it to one project, replace ~/.claude/skills with that project's .claude/skills.
How do you use this skill?
Once installed, send your agent any of these to trigger it:
- Find startup errors from the latest Code OSS development run
- Check MCP errors from the latest Agents app run.
In a client where this skill is available, make a concrete request such as: “Find startup errors from the latest Code OSS development run” or “Check MCP errors from the latest Agents app run.” If a custom user-data directory was used, provide --user-data-dir=<dir> from the launch command; the skill uses <dir>/logs/. Without it, it checks $HOME/.vscode-oss-dev/logs/. Reading the logs requires local shell and filesystem access; the source does not define a client-specific trigger command.
What are this skill's strengths and limitations?
- Covers multiple development processes in Code OSS and the Agents app.
- Explicitly handles the default user-data directory, custom --user-data-dir values, and temporary .build directories.
- Maps startup, extension, agent, MCP, terminal, and network investigations to concrete files.
- Provides directly executable inspection patterns using tail, rg, and find.
- It is scoped to Code OSS or Agents app development-run logs rather than general log analysis.
- It requires local log files plus shell and filesystem access.
- The source provides no standalone installation command, test suite, or cross-platform validation details.
- The temporary console-forwarding workflow requires editing source code and manually restoring it afterward.
How does this skill compare with similar options?
Side by side with related skills; every score comes from the same FSRS standard.
| Skill | FS score | Stars | Last updated | License |
|---|---|---|---|---|
| Code OSS Development Log Viewer this page ✓ Microsoft · Official | 51 · Use with care | ★ 194k | 11d ago | MIT |
| VS Code Integration Test Runner ✓ Microsoft · Official | 52 · Use with care | ★ 194k | 11d ago | MIT |
| OpenChamber UI/API Decoupling Skill | 58 · Recommended | ★ 12k | 11d ago | MIT |
| AionUi Architecture Conventions | 57 · Use with care | ★ 34k | 12d ago | Apache-2.0 |
| Agent Host E2E Test Assistant ✓ Microsoft · Official | 56 · Use with care | ★ 194k | 11d ago | MIT |
How did FollowSkills review this skill?
The skill is primarily local log discovery and reading, with no required network, credential, or destructive operation, and it explicitly requires reverting temporary probes and not committing them. However, logs may contain telemetry, authentication, or proxy data; the document lacks redaction guidance, user confirmation, access isolation, and backup/recovery instructions, so points are deducted.
Paths, log layout, newest-run selection, and use-case-based file mapping are reasonably clear. However, missing directories, permission errors, empty logs, and multi-window ambiguity lack diagnostic handling, command portability is assumed, and the skill has no dedicated key-path tests; the static ceiling therefore requires a conservative deduction.
The audience and scenarios for Code OSS and the Agents app are clear, and common failures map to relevant files. Trigger conditions, non-fit boundaries, distribution differences, Chinese-language usage, and mainland-China environment fit are not specified, and support is limited for non-development user-data layouts, so points are deducted.
The document has front matter, layered headings, tables, directory examples, command examples, and limitation notes; repository material also establishes MIT licensing and Microsoft attribution. It lacks skill versioning, changelog, explicit maintenance owner, update path, and complete installation/troubleshooting notes, so points are deducted.
It provides directly usable file-selection and reading procedures for startup, extension, agent, MCP, terminal, and network investigations, so the core task is plausible. There is no statically verifiable representative output, correctness evidence, or comparison with manual search, and the temporary-probe workflow adds cost, so it does not approach full marks.
The skill text is auditable and gives traceable file names and directory structure. However, there are no tests, fixed sample logs, CI coverage, or third-party reproduction artifacts for this skill; the supplied repository CI does not directly cover its key paths, so only limited points are awarded.
Open a dimension to read why it scored that way
Evidence confidence：Low — Mostly static review, author material or a limited demo; useful for discovery, not high-risk decisions.
See the full review method →