菜单
Use the discover skill to find connections between claims through facet relations and to rank inference and facet candidates using manifold geometry. Covers facet modeling, manifold scoring, relatedness, distance, uncertainty, and retrieval of discovery results.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Use the speculate skill when you need critique-ready hypotheses from new observations: recall context, run the mutation pipeline on a fresh observation, retrieve ranked inference and facet results from an inference/* branch, and present candidate text, provenance, assumptions, uncertainty, and review priority without promoting anything to trusted memory.
Use the infer skill to generate speculative inference candidates from claims, persist them on branch-local inference graphs, and run the full mutation pipeline that coordinates active memory, claims, Terminus persistence, and speculative ranking.
Use the memorize skill to store observations, entities, tasks, and claims into the filesystem-backed active memory. Covers session lifecycle, working-set writes, entity/task card creation, claim extraction, and candidate promotion.
Use the recall skill to retrieve composed context from active memory, trusted Terminus graphs, and optional speculative branches. Also covers journal-backed history of past working-memory mutations.
Use the reflect skill to persist memories, claims, and inference nodes into the TerminusDB temporal graph. Covers branch lifecycle, trusted session writes, temporal recall, and the separation between session and inference branches.
| name | discover |
| description | Use the discover skill to find connections between claims through facet relations and to rank inference and facet candidates using manifold geometry. Covers facet modeling, manifold scoring, relatedness, distance, uncertainty, and retrieval of discovery results. |
Find new connections across knowledge through facets and geometric ranking.
from src.inference.generator import InferenceGenerator
from src.claims.extractor import ClaimExtractor
extractor = ClaimExtractor()
claims = extractor.extract(
text="The auth service failed. The API returned 401 errors.",
source_ref="incident-42",
)
generator = InferenceGenerator()
facets = generator.generate_facet_candidates(
claims,
provenance_commit="journal-event-abc",
source_branch="inference/sess-1",
)
for f in facets:
print(f.facet_type, f.shared_core_claim, f.differences)
from src.manifold_sidecar import ManifoldRankingService, ManifoldRankingRequest
from src.inference.generator import InferenceGenerator
from src.claims.extractor import ClaimExtractor
extractor = ClaimExtractor()
claims = extractor.extract(text="Database writes slowed after index rebuild.")
generator = InferenceGenerator()
candidates = generator.generate_from_claims(
claims,
source_branch="inference/sess-1",
source_commit="evt-1",
session_id="sess-1",
)
facets = generator.generate_facet_candidates(
claims,
provenance_commit="evt-1",
source_branch="inference/sess-1",
)
svc = ManifoldRankingService()
# Rank inference candidates
inference_ranking = svc.rank_inference_candidates(
ManifoldRankingRequest(
branch_name="inference/sess-1",
ranking_mode="inference_candidate_ranking",
seed_context={"session_id": "sess-1"},
candidates=candidates,
)
)
for c in inference_ranking.candidates:
print(c.ranking_score, c.relatedness_score, c.uncertainty)
# Rank facet candidates
facet_ranking = svc.rank_facet_candidates(
ManifoldRankingRequest(
branch_name="inference/sess-1",
ranking_mode="facet_candidate_ranking",
seed_context={"session_id": "sess-1"},
candidates=facets,
)
)
for f in facet_ranking.candidates:
print(f.facet_strength, f.relatedness_score, f.distance_score)
import pathlib
from src.retrieval.composer import RetrievalComposer
from src.terminus.adapter import TerminusMemoryRepository
root = pathlib.Path(".agent-memory")
repo = TerminusMemoryRepository(url="http://localhost:6363")
composer = RetrievalComposer(root, terminus_repo=repo)
result = composer.retrieve(
include_terminus=True,
include_speculative=True,
inference_branch="inference/sess-1",
)
for node in result["speculative_inference"]:
print(node.get("ranking_score"), node.get("text"))
for rel in result["facet_relations"]:
print(rel.get("facet_type"), rel.get("facet_strength"))
| Class | Method | Purpose |
|---|---|---|
InferenceGenerator | generate_facet_candidates() | Produce FacetRelation candidates between claims |
ManifoldRankingService | rank_inference_candidates() | Score and sort InferenceNode candidates |
ManifoldRankingService | rank_facet_candidates() | Score and sort FacetRelation candidates |
ManifoldRankingService | get_current_model_metadata() | Inspect active model, geometry, and feature list |
RetrievalComposer | retrieve() | Compose context including speculative and facet results |
src/inference/models.py — InferenceNode and FacetRelation schemassrc/inference/generator.py — facet generation logicsrc/manifold_sidecar/service.py — manifold ranking servicesrc/manifold_sidecar/models.py — ranking request/response modelssrc/retrieval/composer.py — composed retrieval with discovery layerstests/integration/test_inference_flow.py — integration coveragenotebooks/05_knowledge_discovery_through_facets.ipynb — facet discovery walkthroughnotebooks/06_manifold_mapping_for_discovery.ipynb — manifold ranking walkthroughInferenceGenerator from src/inference/generator.py for facet candidate creation.ManifoldRankingService from src/manifold_sidecar/service.py for all scoring — never hard-code scores outside the sidecar.inference/* or verification/* branch context; a ManifoldBranchContextError is raised otherwise.tests/integration/test_inference_flow.py.