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.
npx skills add vmobifystudio/app-dev-team --skill performance-reviewgit clone --depth 1 https://github.com/vmobifystudio/app-dev-teamWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/skills/vmobifystudio/app-dev-team/performance-review)<a href="https://agentmods.dev/skills/vmobifystudio/app-dev-team/performance-review"><img src="https://agentmods.dev/badge/skills/vmobifystudio/app-dev-team/performance-review/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/vmobifystudio/app-dev-team/performance-review"><img src="https://agentmods.dev/badge/skills/vmobifystudio/app-dev-team/performance-review.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00075 | $0.00712 |
| Opus 5.5 | $0.00030 | $0.00285 |
| Sonnet 5.5 | $0.00015 | $0.00142 |
| Haiku 4.5 | $0.00007 | $0.00071 |
Grade A, and why
performance-review 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 1mo 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.
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.
Performance review
Performance is a budget, and a budget only exists if something can exceed it. The budgets live in
docs/20-architecture.md §8; if that section is empty, that is the finding.
Spawn the platform auditor first and fold its findings in — axiom:swiftui-performance-analyzer,
axiom:swift-performance-analyzer, axiom:memory-auditor, axiom:energy-auditor on Apple
platforms. All external and optional (separate plugins) — missing → record
N/A: <tool> — not installed, walk the list below by hand, never file the absence as a defect.
Measure, then read the diff
A performance finding with no number is an opinion. Every finding names: the metric, the value,
the budget it is measured against, the device or environment, and the build. Where you could not
measure, the verdict is CANNOT EVALUATE — not a guess dressed as a warning.
The dimensions
| Dimension | Budget it spends | What to look for in a diff |
|---|---|---|
| Startup | cold launch to first interaction | work moved into app init, eager singletons, sync I/O on the launch path |
| Frames | 60/120fps, jank-free scroll | layout or allocation per frame, unbounded list without recycling, expensive work in a view body |
| Memory | peak and steady-state | full-size images held, unbounded caches, retain cycles, accumulating observers |
| Battery / energy | background and foreground drain | polling instead of push, wake locks, continuous location, timers that never stop |
| Network | bytes and round trips | N+1 requests, no pagination, no caching headers, retry without backoff |
| Storage | disk footprint and write amplification | writing on every keystroke, unbounded logs, no eviction policy |
| Bundle / binary | download size | a whole library imported for one function, unsplit routes, uncompressed assets |
The rules
- Measure before optimising, and measure the thing users feel. A 40% faster function on a path that runs once at midnight is not a result.
- Regressions are found by comparing to a baseline, so record the baseline value in the ticket. No baseline, no regression detection — that is how these arrive silently.
- A cache is a correctness decision, not a performance one. Every cache added must state its invalidation rule; an unstated one is a stale-data defect waiting.
- Never optimise by removing a bound. Widening a page size, dropping a limit or removing a timeout trades a slow path for an unbounded one.
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.
- 1mo ago First seen · 52 lines · 75 tokens per session scan A 4b04d865106a
performance-review is a skill published in the GitHub repository vmobifystudio/app-dev-team (4 stars, last pushed 4d ago), licensed MIT. It adds 75 tokens to every session and 712 once invoked, about $0.0003 per session on Opus 5.5. 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 2026-09-03.
Other skills, from other repositories
postmortem
Auto-generates a structured postmortem from a completed campaign. Reads the campaign file, telemetry logs, and feature ledger. Produces a documented analysis of what broke, what the safety systems caught, and what patterns emerged. Can also be invoked manually for any incident.
systematic-debugging
4-phase root cause analysis: observe, hypothesize, verify, fix. Enforces investigation before any code changes. Emergency stop after 2 failed fixes. Prevents shotgun debugging and fix cascades.
Systematic debugging
Use when you hit a bug, test failure, or unexpected behavior — find the root cause before changing code.
root-cause
Find the mechanism behind a failure instead of patching its symptom - reproduce first, one variable per experiment with the prediction written before the run, exit by naming the mechanism and pinning it with a failing test. Use for a bug, an unexplained red test, or a failure that will not reproduce.
claude-md-drift-check
Use when detecting drift between CLAUDE.md (or AGENTS.md, the Codex CLI alias) / meta narrative and live repository state. Ten checks: absolute-path resolution, 01-projects/ count claims, issue-reference freshness, session-file existence, command-count sync, session-config-parity (mandatory template keys = error…
debug
Use when encountering any bug, test failure, build break, or unexpected behavior — runs a 4-phase systematic debugging process before proposing any fix. Iron Law: NO FIXES WITHOUT ROOT CAUSE INVESTIGATION FIRST. Produces a .orchestrator/debug/ artifact the fixer agent must reference.