api-platform-performance

A set of guidance for improving API Platform 4.3 applications that use Doctrine and Symfony. API Platform is a framework for building web APIs; Doctrine is a database access library, and Symfony is a PHP framework.

In plain words
What is it for?
It is for diagnosing eager loading and N+1 queries, shaping list results into DTOs, configuring metadata and HTTP caching, and processing large collections more safely.
Why use it?
It addresses slow API responses, excessive database queries, high memory use, cache invalidation problems, and oversized result processing.

Skill for Claude CodeCodex

Part of the gerard plugin — 25 skills, 25 commands, 7 agents, 1 hook shipped together

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add skills/gerard-labs/superpowers-api-platform/api-platform-performance
Any agent
npx skills add gerard-labs/superpowers-api-platform --skill api-platform-performance
Clone the repo
git clone --depth 1 https://github.com/gerard-labs/superpowers-api-platform

Made for: Claude Code, Codex.

Or install gerard, the plugin that ships this one along with the rest of its 25 skills, 25 commands, 7 agents, 1 hook.

Per session 237 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 777 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00237 $0.00777
Opus 5 $0.00118 $0.00388
Sonnet 5 $0.00047 $0.00155
Haiku 4.5 $0.00024 $0.00078

Measured 2d ago against content hash 4f46e535c21d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

api-platform-performance scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 2d ago.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/api-platform-performance/SKILL.md · 52 lines

How it starts

The opening of the file, as written. The whole thing — 52 lines — stays where its author put it; the contents beside it link to each section on GitHub.

API Platform 4.3 — Performance

Use when

  • The API is slower than expected, especially on list endpoints.
  • A max_joins: 30 exception fires because the entity has many relations.
  • Memory blows up on large collections.
  • p99 latency is above ~50 ms for simple routes.
  • Cache invalidation cannot keep up with writes.

Default workflow

  1. Detect the force_eager: true trap. Switch to false and replace with targeted join fetches in repositories.
  2. Profile with the Symfony profiler in dev — look for N+1 patterns.
  3. For list views: project a DTO via DQL NEW App\Dto\PostListItem(...) instead of hydrating the full entity.
  4. Enable APCu for resource metadata cache in prod.
  5. Adopt FrankenPHP worker mode for high-traffic APIs (stateful kernel between requests).
  6. Add HTTP cache headers per operation and cacheTags for invalidation.

Guardrails

  • force_eager: true is the 4.x default but rarely what you want for an entity with many relations. Set it to false and use targeted addSelect() join fetches per query.
  • #[ApiProperty(fetchEager: true)] for specific fields only — never blanket eager loading.
  • Never iterate without clear() on million-row collections — memory will explode.
  • HTTP cache + write paths must agree on invalidation — purge cacheTags on every write.
  • FrankenPHP worker mode requires worker-safe code (no static state leaks).

Progressive disclosure

  • SKILL.md covers posture and rules.
  • reference.md carries the full patterns: fetch modes, eager loading trap, join fetch recipe, DTO projection, batch iteration, HINT_READ_ONLY, APCu metadata cache, FrankenPHP worker, HTTP cache headers, cache tags + Varnish PURGE, dynamic Last-Modified listener.

Output contract

  • eager_loading.force_eager: false confirmed (or documented exception).
  • Repository methods using targeted addSelect() for list endpoints.
  • DTO projections via DQL NEW for large lists.
  • APCu metadata cache configured in prod.
  • HTTP cache configured per operation (where applicable).
  • cacheTags purged on writes.
  • p99 latency measured before/after.

Read the full file on GitHub · 52 lines

Files

What ships with it

1 file beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 2d ago First seen · 52 lines · 0 tokens per session scan A 4f46e535c21d

Subscribe to this mod's changes

api-platform-performance is a skill published in the GitHub repository gerard-labs/superpowers-api-platform (2 stars, last pushed 3mo ago), licensed MIT. It adds 237 tokens to every session and 777 once invoked, about $0.0012 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.