imMamdouhaboammar/anti-slop-engineering / skills/context-window-and-token-economics-architect
context-window-and-token-economics-architect
VerifiedReview LLM context selection, retrieval evidence, prompt cache boundaries, time to first token, and structured output contracts. Use for changes that affect what an LLM sees, reuses, or returns.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 4.7 KB |
| examples/paired-workflow.md | markdown | 1.4 KB |
| examples/refund_workflow.py | script | 1.8 KB |
examples/paired-workflow.md
View on GitHub# Refund policy retrieval comparison
Return the current refund deadline as an integer and cite the policy passage by source ID. If current evidence is absent or unauthorized, return `no_evidence`.
| Input | Baseline | Candidate |
| --- | --- | --- |
| Retrieval | Entire authorized policy and receipt guide | Authorized refund passage with source ID and revision |
| Cache identity | Tenant and query | Tenant, document revision, authorization scope, and output schema revision |
| Offline rendered input | Refund passage, receipt instructions, and unrelated policy | Refund passage with its provenance |
| Offline budget | 20 whitespace-delimited units | The same 20-unit budget |
The service supplies the current revision from trusted document metadata and the access scope from its authorization check. Prompt text supplies neither value. The fixture checks the answer against the authorized source, then exercises a revision change from 30 to 14 days, access revocation, a second tenant without evidence, and a schema revision. It reports input size with a whitespace counter for reproducibility. Use the selected API's tokenizer or returned usage for real token counts and billing.
For a live comparison, fix the model and schema and stream both variants over the same sanitized, labeled task set. Compare cold and warm runs separately. Set the latency budget against the service objective and require every refund deadline in this example to match its cited revision.
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 472k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 323k | Updated | Claude CodeCodex | Verified | Install |