agentleFS
vllm-projectvllm-project/agentic-api/.agents/skills/executor-architecture-review/SKILL.md

executor-architecture-review

Skill346 starsChanged 6 days agoApache-2.0View on GitHub

Use when planning or reviewing agentic-api changes that affect executor streaming, upstream SSE handling, output-item lifecycle, stream delivery, backpressure, or worker placement.

The file

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 2 are distinctive, and 3 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 3

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Use the PR review skill instead for a broad pull-request review that does not involve these boundaries.

    line 11
  2. Very specific:

    Read AGENTS.md and TERMINOLOGY.md completely.

    line 17
  3. Very specific:

    Read ARCHITECTURE.md, especially "Target streaming pipeline and ownership boundaries" and the affected module sections.

    line 18
  4. Very specific:

    Check output-item lifecycle invariants: validated output_index, stable item ID and kind, typed active state, detectable index …

    line 28
  5. Distinctive:

    Check concurrency boundaries.

    line 30

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Executor Architecture Review

Perform a read-only review that protects the executor's streaming ownership boundaries. Do not edit files, create or update issues, post comments, commit, or push.

Use the PR review skill instead for a broad pull-request review that does not involve these boundaries.

Authoritative inputs

Apply the shared engineering preparation first; reuse any completed discovery.

  1. Read AGENTS.md and TERMINOLOGY.md completely.
  2. Read ARCHITECTURE.md, especially "Target streaming pipeline and ownership boundaries" and the affected module sections. Treat it as the source of truth; this skill does not replace it.
  3. Resolve the exact task, diff, or pull request and inspect the complete changed path with relevant surrounding code.

Review workflow

  1. Map each changed responsibility to exactly one owner: inference transport, event normalization, synchronous ingestion, stream relay, or engine orchestration. Flag logic that has no owner or appears in multiple stages.
  2. Trace one representative streaming event from upstream bytes through client delivery. Check that each stage consumes the previous stage's typed result instead of re-parsing or reconstructing its state.
  3. Check output-item lifecycle invariants: validated output_index, stable item ID and kind, typed active state, detectable index reuse/duplicate completion, exhaustive kind handling, and consuming terminal finalization.
  4. Check concurrency boundaries. Every channel needs entry and memory bounds, full/disconnect behavior, cancellation propagation, and join/error handling. Treat spawn_blocking, worker placement, and proposed capacities as claims requiring representative measurements, not defaults.
  5. Match tests to the changed owner. Consider malformed and out-of-order semantic events, ID/index mismatches, duplicate completion, incomplete terminal state, tiny capacities, oversized items, slow consumers, cancellation, and task failure when applicable. Do not demand unrelated cases.

Findings

Report only actionable correctness, regression, performance, or missing-test issues. Verify each finding against the actual path and use this form:

[P1] path/to/file.rs:123 — short title
Boundary: <owner and violated responsibility>
Evidence: <specific path or scenario>
Impact: <observable failure>
Fix: <concise direction that restores one owner and one path>

Order findings from P0 to P3. If none remain, say so explicitly. End with a short ownership map, tests or benchmarks inspected or run, and any environment limitations.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.