What does this skill do, and when should you use it?
This is the Agent Skill for the sessions architecture of the Agents window in microsoft/vscode, located at .github/skills/sessions/SKILL.md. It covers the agents-first application, layering, folder structure, chat widgets, menus, contributions, entry points, and development guidance. Before changing code under src/vs/sessions/, it requires the agent to read the coding and source-organization guidelines plus the relevant specification, and to keep that specification synchronized with implementation changes. It is intended for agents maintaining VS Code Agents window code, not for general editor usage or unrelated coding tasks.
It directs an agent to read the two required documents under .github/instructions/ and the relevant layer, layout, sessions, sessions-list, mobile, or AI-customization specification under src/vs/sessions/. It requires adherence to the internal core → services → contrib → providers layering, the shared Sessions menu registry, observable session-state propagation, and the appropriate sessions.*.main.ts entry point for new contributions. It also supplies concrete implementation pitfalls and repair rules covering layout restoration, session-versus-chat modeling, provider isolation, mobile layouts, editor tabs, permissions, subagent chats, and asynchronous state management. Its output is source-code work that follows these constraints together with synchronized specification updates; the supplied source does not define an executable script or standalone command.
- A VS Code developer adding an Agents window feature under src/vs/sessions/ uses it to identify the required guides and specifications.
- A maintainer fixing session switching, layout restoration, sidebar, or auxiliary-bar behavior uses it to check state and grid-layout constraints.
- A developer changing session providers, chat data flow, or multi-chat support uses it to avoid provider leaks and incorrect model access.
- An engineer working on the mobile Agents window uses it for viewport, drawer-sidebar, and view-gating rules.
- A contributor changing Agents window menus, editor integration, or contribution loading uses it to verify registration and entry-point requirements.
How do you install this skill?
- The skill depends on many sessions specification files not included in the supplied material; verify that those paths and contents exist and are synchronized at the assessed revision.
- Do not treat repository-wide CI or Microsoft provenance as independent validation of this skill; no code or tests were executed for this review.
- The large rule set lacks explicit prioritization and conflict-resolution guidance, so complex changes may still require human judgment.
- Local filesystem
The Skill is located at .github/skills/sessions/SKILL.md in the repository. Place this file from microsoft/vscode in a compatible Agent Skills directory while preserving its path and context; the supplied sources do not document a client-specific installation command or directory.
tmp="$(mktemp -d)"
git clone --depth 1 https://github.com/microsoft/vscode.git "$tmp"
mkdir -p ~/.claude/skills
cp -R "$tmp/.github/skills/sessions" ~/.claude/skills/
rm -rf "$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:
- Modify the sessions list under src/vs/sessions/, first read the coding guidelines, source-organization guide, and SESSIONS_LIST specification, then update the specification with the implementation.
Trigger it when implementing or fixing an Agents window feature, for example: "Modify the sessions list under src/vs/sessions/, first read the coding guidelines, source-organization guide, and SESSIONS_LIST specification, then update the specification with the implementation." Before making changes, follow the Skill's required reading order and consult the specification for the affected area.
What are this skill's strengths and limitations?
- Covers engineering constraints across Agents window layering, layout, session models, and contribution entry points.
- Includes concrete rules for recurring implementation failures, such as avoiding delay-based race fixes, preserving provider isolation, and using observable state.
- Maps directly to VS Code source directories and specification files, making it useful for repository maintenance.
- The parent repository is MIT-licensed and developed by Microsoft together with the community.
- Its scope is limited to the VS Code Agents window and does not represent the capabilities of the repository's other 60 Skills.
- It assumes access to and familiarity with the VS Code source tree and its specifications.
- The supplied SKILL.md is truncated, so any omitted content remains unknown.
- The sources do not provide a standalone installation command, test suite, or runtime dependency list for this Skill.
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 |
|---|---|---|---|---|
| VS Code Agents Window Sessions this page ✓ Microsoft · Official | 56 · Use with care | ★ 194k | 11d ago | MIT |
| Cost Mode | 54 · Use with care | ★ 38 | 11d ago | MIT |
| Component Fixtures for Screenshot Testing ✓ Microsoft · Official | 53 · Use with care | ★ 194k | 11d ago | MIT |
| Search View Results Reader ✓ Microsoft · Official | 43 · Not recommended | ★ 194k | 11d ago | MIT |
| Angular Developer Skill | 63 · Recommended | ★ 685 | 11d ago | — |
How did FollowSkills review this skill?
The skill narrowly scopes work to src/vs/sessions/ and requires reading governing specifications; it also stresses provider isolation, correct denial semantics, and deterministic state handling. It does not specify user confirmation, sensitive-data flows, rollback, or external-side-effect controls, so points are deducted.
The guidance is detailed and mostly internally consistent, covering entry points, state, async behavior, and abnormal cases. However, this is a static review with no skill-specific reproduction, execution results, or unified diagnostic procedure; the score remains below the static ceiling.
The audience and trigger scenario—implementing or fixing the Agents window—are clear, with desktop, mobile, and provider boundaries described. Non-fit tasks, formal inputs/outputs, exclusion triggers, Chinese-language support, and mainland-China network fit are not defined, so points are deducted.
The document has useful progressive structure through prerequisites, a specification table, and common pitfalls; repository context also supplies MIT licensing, Microsoft attribution, and maintenance signals. The skill itself lacks versioning, changelog, dependency/install guidance, FAQs, and explicit maintenance ownership.
The rules can directly guide Agents-window code changes and identify many concrete failure patterns with corrective direction. Still, there is no output template, representative completed result, or skill-specific verification evidence, so direct usability and marginal benefit are only partially established.
The skill cites concrete specifications, entry points, and implementation concepts, while the repository includes CI and test infrastructure. The supplied files do not demonstrate that those tests cover this skill’s key paths or provide revision-specific execution results, so the score remains conservative.
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 →