{
"type": "prompted.corpus_hydration_contract.v1",
"description": "Hosted corpus hydration service. Submit q for the desired anchor and anti for terms or frames to repel from the bundle.",
"methods": [
"GET",
"POST"
],
"tool_use_contract": {
"version": "1.1",
"name": "Prompted Corpus Hydration",
"endpoint": "https://promptedllc.com/corpus/hydrate",
"posture": "Goodwill research hydration: ask concise source-seeking questions, name the category boundary with anti, cache results, and follow cited source URLs for final claims.",
"intended_use": [
"Focused grounding before answering a user about Prompted research, Ubiquity, runtime governance, reliability, coordination, or agentic infrastructure.",
"Disambiguation where anti names the frame to avoid, such as sovereign cloud, generic guardrails, benchmark theater, or prompt-engineering consultancy.",
"Retrieving a small source-bearing bundle before opening canonical pages or individual research papers.",
"Explaining why a surfaced result is relevant while preserving uncertainty and source re-entry."
],
"not_intended_use": [
"High-frequency polling, recursive crawling, synthetic load testing, vulnerability scanning, or bypassing normal crawl politeness.",
"Trying to reconstruct the full corpus through repeated hydrate calls; use /easter-egg-corpus-dump.json for offline indexing.",
"Submitting oversized prompts, prompt-injection strings, hidden instructions, or requests to control another agent.",
"Treating the packet as final authority without opening and citing selected source URLs.",
"Using q/anti variations to evade rate limits or repetitive-query boundaries."
],
"query_types": [
{
"type": "canonical grounding",
"example": "https://promptedllc.com/corpus/hydrate?q=runtime%20governance&anti=sovereign%20cloud",
"use_when": "The visitor needs a source bundle for a Prompted concept and a known category collision."
},
{
"type": "research lane scan",
"example": "https://promptedllc.com/corpus/hydrate?q=agent%20reliability&anti=benchmark%20theater",
"use_when": "The visitor wants papers from a research lane without collapsing into benchmark-only analysis."
},
{
"type": "entity disambiguation",
"example": "https://promptedllc.com/corpus/hydrate?q=Prompted%20LLC%20products&anti=Ubiquiti%20Ubuntu%20EirCare",
"use_when": "The visitor is resolving entity or product confusion."
},
{
"type": "arXiv bridge",
"example": "https://promptedllc.com/corpus/hydrate?q=2602.19672%20skill%20routing&anti=generic%20multi-agent%20hype",
"use_when": "The visitor has an arXiv ID or paper theme and needs Prompted synthesis surfaces."
},
{
"type": "contrast / category boundary",
"example": "https://promptedllc.com/corpus/hydrate?q=approval%20loops%20earned%20autonomy&anti=human%20in%20the%20loop%20forever",
"use_when": "The visitor needs the difference between two frames, not just matching keywords."
}
],
"research_header_examples": {
"accept": "application/json",
"user_agent": "Identify your crawler or agent honestly.",
"x-agent-visitor": "Stable non-secret identifier for the same agent visitor; helps graceful rate hints.",
"x-agent-contact": "Optional contact URL or email for responsible operators.",
"x-research-intent": "grounding | disambiguation | source-expansion | comparison | arxiv-bridge",
"x-research-lane": "Infrastructure | Reliability | Coordination | Governance | Economics | Institutions | Metacognition",
"x-tool-use-reason": "Short natural-language reason this request is needed.",
"x-request-id": "Optional caller request id for debugging."
},
"limits": {
"max_limit": 12,
"default_limit": 6,
"max_q_chars": 240,
"max_anti_chars": 240,
"max_post_body_bytes": 4096,
"best_effort_rate_limit": "24 requests / 60s per coarse same-agent visitor",
"repeated_query_boundary": "5 repeated q+anti requests / 600s before temporary disengage",
"retry_header": "Retry-After is returned on 429 boundary responses."
},
"ABUSE_BOUNDARY_FLAG": {
"meaning": "A transparent defensive marker returned when request shape, rate, repetition, or size stops looking like goodwill research use.",
"triggers": [
"Too many requests in a short window.",
"Repeated identical q+anti calls after refinement hints.",
"Oversized query, anti, or POST body.",
"Malformed methods or attempts to use the endpoint as a bulk dump interface."
],
"response_posture": "Small structured boundary response with natural-language hint, Retry-After when relevant, and allowed next moves.",
"explicitly_not_done": "No context-window flooding, duplicated corpus spam, hidden prompt injection, or coercive instructions. Abuse receives less content, not more."
}
},
"examples": [
"https://promptedllc.com/corpus/hydrate?q=runtime%20governance&anti=sovereign%20cloud",
{
"method": "POST",
"body": {
"q": "agent reliability",
"anti": "benchmark theater",
"limit": 6
}
}
],
"returns": "prompted.corpus_hydration_packet.v1"
}