Core Concepts

Core Concepts

PIC uses a small set of operational concepts. These terms are the most important for first-time users and agents.

Candidate

A candidate is material that may be useful but is not verified yet.

Examples include an LLM answer, an agent message, a repository finding, a web excerpt, a test claim, or a proposed reusable abstraction.

PIC treats candidates as data to inspect, not as facts to trust.

Evidence

Evidence is something a checker can inspect.

Useful evidence may include hashes, schemas, signed attestations, verifier reports, test logs, replayable artifacts, provenance records, or typed trace records.

Metadata such as "passed", "approved", "important", or "high priority" is not evidence by itself.

Proof Obligation

A proof obligation is a check that must still be completed.

PIC records obligations instead of hiding them. An obligation can be routed to a verifier, deferred, or preserved as residual work.

Residual Ledger

A residual ledger is a structured list of unresolved debt.

Residuals can include missing evidence, stale data, external-domain limits, verifier backlog, unsafe packet status, resource mismatch, or unverified physical assumptions.

Keeping residuals visible is safer than rewriting uncertainty as success.

Verifier Route

A verifier route is a named path for checking a specific kind of obligation.

It might point to a schema check, evidence verifier, identity check, trace checker, snapshot verifier, domain adapter, or runtime route.

A route is not evidence by itself. It is only the path where evidence can be checked.

Phase Ecology Lab

Phase Ecology Lab is the v0.5.0 and later local workbench for looking at several reports together.

It builds phase windows, effective packet graphs, closure diagnostics, execution-available path diagnostics, and certificate candidates. It is diagnostic-only. It does not execute report text or turn raw packet volume into progress.

Effective Packet Graph

An effective packet graph shows packet candidates, support edges, evidence references, blockers, and residuals.

Plain-language equivalent: a map of which work items support which other work items, and why some support is still weak.

Reusable Abstraction Capital

Reusable abstraction capital is a scoped reusable pattern, trace, packet, or idea that can reduce future work.

PIC's ALT layer checks whether reuse is still beneficial after validation, transport, lifecycle, hazard, maintenance, and residual costs.

Raw agent output, web content, or trace material is only a candidate until these checks pass.

Settlement

settled=true means the relevant scoped finite obligations were discharged.

settled=false is common and useful. It means the report preserved unresolved work instead of pretending the work is complete.

Operational Usability

workflow_usable=true means the compact agent workflow can guide next safe actions.

operationally_usable=true means the report can guide routing under the selected profile.

CCR Interop

CCR interop is the v0.6.0 data boundary between PIC and Collective Capability Runtime.

PIC can emit ccr.task.v0.1 and ccr.residual.v0.1 JSONL records. Those records are candidate scheduling inputs. They do not execute commands and do not settle downstream runtime work.

TRC Operation Readiness

TRC operation readiness is a finite check over a declared trace normal form.

operation_ready=true means the trace includes scoped authority, resources, rollback or escrow, witnesses, schedule, and tolerance data. It does not mean the operation was executed or that the external result is true.

It does not mean every claim is true, every obligation is settled, or every action is safe to execute.

Live Candidate Intake

Live candidate intake means PIC can fetch or read an explicitly supplied source under byte limits, content-type checks, private-network rejection, provenance recording, and residual preservation.

Live intake is not background crawling, general browsing, shell execution, or automatic trust. The result is still a candidate until downstream checks accept it.