agentleFS
hugobatistahugobatista/agent-skills/skills/review/saas-anti-pattern-review/SKILL.md

saas-anti-pattern-review

Skill1 starsChanged 58 days agoMITView on GitHub

Investigate code or an issue for SaaS-unfriendly patterns.

The file

We split this skill into the rules it gives an agent

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

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 2

    Very specific

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

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Do multiple components assign websocket.onmessage = handler (collision-prone)?

    line 63
  2. Very specific:

    Should use addEventListener('message', handler) instead.

    line 64

≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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

Investigate code or an issue for SaaS-unfriendly patterns. Use when the user asks to review something for SaaS readiness or to check if a pattern scales horizontally.

Trigger phrases

"review for SaaS", "SaaS readiness", "SaaS anti-pattern", "SaaS-friendly", "is this SaaS-ready", "check this for SaaS", "will this scale for SaaS"

Investigation checklist

For each file or area being reviewed, check these patterns:

1. State locality

  • Is in-memory state (plain dict, module-level variable, @lru_cache singleton) used for:
    • WebSocket connections?
    • MFA / login / lockout state?
    • Rate limiting counters?
    • OAuth state or PKCE verifiers?
    • Failed login attempt tracking?
  • In-memory state is invisible to other instances. Requires Redis or a shared KV store for multi-instance deployments.

2. Blocking request patterns

  • Does any endpoint use asyncio.to_thread or run_in_executor for long-running sync work?
  • Is there a future.result(timeout=N) blocking > a few seconds?
  • Are background tasks submitted via BackgroundTasks (not durable — lost on restart)?
  • Long blocking requests exhaust the thread/connection pool. Should be async job queues.

3. WebSocket dependency

  • Is WebSocket the only delivery mechanism for a critical user-facing event?
  • Is there a fallback (polling, job status endpoint) when WS is unavailable?
  • WS should be an optimization, never a requirement for core flows.

4. Scheduler / job duplication

  • Are periodic tasks started in the API process (APScheduler, startup hooks)?
  • Is there a distributed lock or leader election?
  • Without coordination, each replica runs the same job — duplicates imports, notifications, API calls.

5. Filesystem dependency

  • Are uploads, media, temp files written to local paths?
  • Is there an S3 / object-storage abstraction already?
  • Local filesystem prevents horizontal scaling. Needs shared or object storage.

6. Database pool pressure

  • What are the pool size and max overflow settings?
  • Are COUNT queries loading full rowsets (len(results) instead of COUNT(*))?
  • Pool exhaustion is a common failure mode when scaling from single-instance to multi-worker.

7. Request cancellation

  • Does the frontend set client-side timeouts on HTTP requests (AbortController)?
  • Can the backend cancel in-flight work when the client disconnects?
  • Without timeouts, a hanging request occupies a thread/connection indefinitely.

8. Credentials in URLs

  • Are access tokens passed in WebSocket URLs (ws?token=...)?
  • Are API keys accepted via query parameter?
  • URLs end up in logs, browser history, referrer headers, support screenshots.

9. Multiple WebSocket consumers (frontend)

  • Do multiple components assign websocket.onmessage = handler (collision-prone)?
  • Should use addEventListener('message', handler) instead.

Output format

Present findings as a structured list:

Source

{issue / PR / code path}

Blockers found

{N}. [severity] Title — file:line

  • Current behavior: ...
  • Why it fails in SaaS: ...
  • Migration path: ...

Not affected

  • {patterns checked and found clear}

Cross-reference notes

  • {any connections to existing architecture docs or previous findings}

Severity legend

  • 🔴 Critical — blocks core SaaS capabilities (horizontal scaling, zero-downtime)
  • 🟡 Moderate — affects UX, maintainability, or observability
  • ⚪ Low — nice to fix but not blocking

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.