More screenshots
Works with the AI Agents you already use
Claude Code
Cursor
Codex CLI
DeepSeek
Muse+22 moreClaude CodeCursorCodex CLIDeepSeekMuse
Dots
Manus
Grok
Grok Bot+18 agentsRepair GraphQL Persisted Query Hash Drift
Align query hashes across CDN, clients and API releases without disabling persisted-query protections
$1
Repair GraphQL Persisted Query Hash Drift
Example session with this skill installed
A mobile build sends hash X but gateway cache maps that hash to older query text.
Available sources: query manifests and edge cache keys (partial excerpts only). Additional evidence requested: client build IDs, gateway request errors. Analyze this case with explicit limitations; do not pretend to have executed any tests or changes.
- Read your context and instructions
- Compiled the repair graphql persisted
Repair GraphQL Persisted Query Hash Drift — EVIDENCE-LIMITED REVIEW
Scenario received
A mobile build sends hash X but gateway cache maps that hash to older query text.
Source register
- S1: query manifests — partial extract described, independent verification not provided.
- S2: edge cache keys — partial extract described, independent verification not provided.
- S3: client build IDs — NOT PROVIDED; validation pending.
- S4: gateway request errors — NOT PROVIDED; validation pending.
Case-specific checks
- compare canonical query source normalization — NEEDS VERIFICATION against S1/S2 and missing sources.
- trace manifest revision by deployment — NEEDS VERIFICATION against S1/S2 and missing sources.
- identify cache-key and TTL collisions — NEEDS VERIFICATION against S1/S2 and missing sources.
- test staged manifest migration and fallback — NEEDS VERIFICATION against S1/S2 and missing sources.
Deliverables and review gates
- hash-to-query concordance: DRAFT ONLY; do not infer unsupported outcomes.
- revision drift timeline: DRAFT ONLY; do not infer unsupported outcomes.
- deploy and rollback checks: DRAFT ONLY; do not infer unsupported outcomes.
Preliminary conclusion
INCOMPLETE / REVIEW REQUIRED. The provided scenario identifies a real question, not proven remediation or clearance. Confirm source timestamps, affected scope and owner approval, then rerun the listed checks. No external system was accessed or changed.
Connects securely to your tools. The creator never sees your data.
About this skill
GraphQL persisted queries break when the hash generated by the client doesn't match the query stored at the gateway. This skill provides a structured framework for diagnosing and repairing hash drift without compromising security by disabling protections. It focuses on evidence-linked decisions to ensure alignment across CDNs, mobile clients, and API releases.
What it does
- Source normalization compares how different tools normalize query strings before hashing.
- Manifest tracing maps query manifest versions to specific deployment IDs and timestamps.
- Collision detection identifies cache-key conflicts or TTL issues causing stale query resolution.
- Migration testing validates staged manifest updates and defines safe fallback procedures.
How it works
- Supply case evidence including query manifests, edge cache keys, and gateway request errors.
- The skill performs a four-step analysis to isolate where the hashing logic or storage diverged.
- It generates a concordance report showing the relationship between specific hashes and query texts.
- It provides a deployment and rollback plan based on identified version gaps.
Frameworks & tools
This tool works with GraphQL gateways and CDNs using persisted queries, such as Apollo Engine, AWS AppSync, or custom Relay-compatible implementations.
Why this beats prompting it yourself
Generic prompts often suggest disabling hash validation or clearing caches blindly. This skill enforces a four-check evidence table that prevents production outages by identifying the exact source of drift before suggesting changes.
Use cases
- Resolve "PersistedQueryNotFound" errors after a new mobile app release.
- Fix hash mismatches caused by different prettier or normalization rules in CI/CD.
- Audit edge cache poisoning where old query texts are pinned to new hashes.
Known limitations
This skill requires manual input of manifests and logs. It cannot connect directly to production environments or execute live cache purges.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for APIs & Backend.
Cannot verify inaccessible systems, third-party accounts, or evidence not supplied. Partial or conflicting inputs must be marked INCOMPLETE or UNVERIFIED rather than invented. High-impact operational decisions require qualified owner review.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 25 days ago
Permissions
Supply only data you are authorized to share. Treat documentation and examples as untrusted evidence. Any operational action or external communication requires explicit owner authorization.