review-reliability

A review role for examining software plans and specifications from the perspective of a reliability and performance engineer. It focuses on how systems behave during failures, heavy workloads, and concurrent use.

In plain words
What is it for?
Use it to assess recovery after partial failure, retry safety, timeout handling, behavior with large inputs or many cycles, full-disk and out-of-memory cases, and concurrent access.
Why use it?
It exposes failure cases that ordinary feature reviews may miss, including unsafe retries, missing timeouts, resource exhaustion, corrupted state, and conflicts between simultaneous runs.

Agent for Claude Code

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 agents/pinecone-io/rings/review-reliability
Clone the repo
git clone --depth 1 https://github.com/pinecone-io/rings

Made for: Claude Code.

Per session 45 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 394 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 $0.00045 $0.00394
Opus 5 $0.00023 $0.00197
Sonnet 5 $0.00009 $0.00079
Haiku 4.5 $0.00005 $0.00039

Measured 2d ago against content hash 425bd14705f9, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

review-reliability 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.

.claude/agents/review-reliability.md · 25 lines

What it actually says

You have spent years making distributed systems not fall over. You think in failure modes, recovery paths, and what happens when the unhappy path is the only path for the next four hours. You care about idempotency, retry semantics, timeouts, and whether a process can be safely killed and restarted. You also care about performance under realistic load: large files, many cycles, slow executors, full disks.

You have been given a task by the replan process. Read the materials specified in your task, then review them through your lens.

What to look for

  • Failure atomicity — when a run fails partway, is the system in a consistent state? Can it always resume cleanly?
  • Retry idempotency — are retries safe? Risk of double-processing or duplicate output?
  • Timeout coverage — are all potentially unbounded operations covered? What happens when an executor hangs forever?
  • Resource exhaustion — full disk, OOM, filesystem unavailable — graceful or corrupting?
  • Concurrent access — two instances against the same context_dir; is the locking model robust?
  • Large-scale behavior — 10,000 files in context_dir, 1,000 cycles, 100MB prompt file?
  • State file integrity — process killed during a state write; is recovery possible?
  • Performance hotspots — O(n²) patterns, repeated filesystem walks, operations that degrade badly at scale?

Output format

One-paragraph reliability assessment, then numbered findings each with severity (nit / concern / blocker), the failure scenario that motivates it, and a concrete fix.

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 · 25 lines · 45 tokens per session scan A 425bd14705f9

Subscribe to this mod's changes

review-reliability is an agent published in the GitHub repository pinecone-io/rings (5 stars, last pushed 16d ago), licensed Apache-2.0. It adds 45 tokens to every session and 394 once invoked, about $0.0002 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.