Guides selective web search and citations for high-value claims like versions, APIs, security, and performance to optimize token usage without sacrificing accuracy.
From conservenpx claudepluginhub athola/claude-night-market --plugin conserveThis skill uses the workspace's default tool permissions.
Designs and optimizes AI agent action spaces, tool definitions, observation formats, error recovery, and context for higher task completion rates.
Enables AI agents to execute x402 payments with per-task budgets, spending controls, and non-custodial wallets via MCP tools. Use when agents pay for APIs, services, or other agents.
Compares coding agents like Claude Code and Aider on custom YAML-defined codebase tasks using git worktrees, measuring pass rate, cost, time, and consistency.
Intelligent sourcing that requires citations only when the cost is justified by the value of verification.
Full sourcing is prohibitively expensive (10-16x token increase). Smart sourcing targets high-value claims where verification materially improves accuracy.
| Claim Type | Example | Why Source |
|---|---|---|
| Version numbers | "Python 3.12 added..." | Versions change, easy to verify |
| Performance claims | "30% faster than..." | Quantitative claims need evidence |
| Security recommendations | "Use bcrypt for..." | Security advice must be current |
| API specifications | "The function accepts..." | APIs change between versions |
| Release dates | "Released in Q4 2025" | Factual, verifiable |
| Pricing/limits | "Free tier allows 1000 requests" | Business terms change |
| Deprecated features | "X was removed in version Y" | Breaking changes need verification |
| Claim Type | Example | Why No Source |
|---|---|---|
| General concepts | "Async improves concurrency" | Foundational knowledge |
| Code examples | Demonstrative snippets | Illustrative, not factual claims |
| Opinion/preference | "Consider using..." | Clearly framed as suggestion |
| Common knowledge | "Git tracks changes" | Universal understanding |
| Logical derivations | "Therefore, X implies Y" | Reasoning, not factual claim |
| Syntax | "Use def for functions" | Language documentation is stable |
Is this a factual claim?
├── No → No source needed
└── Yes → Would being wrong cause harm?
├── No (minor inconvenience) → No source needed
└── Yes → Is verification cost-effective?
├── No (>5k tokens for low-value claim) → State uncertainty
└── Yes → Perform web search, cite source
When sourcing is warranted:
Python 3.12 introduced `type` statement for type aliases[1].
Sources:
1. [PEP 695](https://peps.python.org/pep-0695/)
When sourcing isn't cost-effective but uncertainty exists:
Python 3.12 introduced the `type` statement (verify in release notes for your version).
| Confidence | Action |
|---|---|
| High (stable facts) | State directly |
| Medium (may have changed) | Add verification note |
| Low (uncertain) | Source or acknowledge uncertainty |
| Approach | Token Cost | Use Case |
|---|---|---|
| No sourcing | Baseline | General knowledge |
| Uncertainty marker | +5-10 tokens | Medium confidence |
| Single source | +2-5k tokens | High-value claim |
| Multiple sources | +10-20k tokens | Research tasks |
User: What's new in FastAPI 0.109?
Response: FastAPI 0.109 added Pydantic v2 support as default[1] and
improved OpenAPI schema generation[2].
Sources:
1. [FastAPI Release Notes](https://fastapi.tiangolo.com/release-notes/)
2. [GitHub Release 0.109.0](https://github.com/tiangolo/fastapi/releases/tag/0.109.0)
User: How does async/await work?
Response: Async/await enables non-blocking I/O by allowing coroutines
to yield control while waiting for operations. The event loop manages
execution, resuming coroutines when their awaited operations complete.
The connection pool default is typically 10 connections (verify in your
database driver documentation as this varies by library and version).
Works with:
conserve:response-compression - Keeps sourced responses conciseconserve:token-conservation - Weighs source cost vs valuememory-palace:research - Full sourcing for knowledge corpusEscalate to full sourcing (accept high token cost) for:
For these cases, use memory-palace:research workflow which is designed for comprehensive sourcing.