PIOS · Capability Library

Durable abilities, independent from any one app.

The library preserves capability meaning extracted from PIOS design sources. Inclusion does not make a capability mandatory for Core, prove an implementation, or establish compatibility.

117 capability CRDs · 31 supporting documents · source 1ea8b8ea8b01
draft · v0.1 · proposed

Retain governed content

Preserve supplied content as a durable governed source object.

draft · v0.1 · proposed

Derive content from a retained source

Produce a traceable derivative without erasing source identity.

draft · v0.1 · proposed

Allocate retained content to a context

Grant and revoke bounded eligibility for content use.

draft · v0.1 · proposed

Assemble governed retrieval context

Select a bounded, explainable context package before consumption.

draft · v0.1 · proposed

Consolidate conversation history

Create a provenance-linked compact representation of a conversation range.

draft · v0.1 · proposed

Decide a governed proposal

Preserve a suggestion through an authorized decision lifecycle.

draft · v0.1 · proposed

Establish a standing rule

Create bounded reusable authority through explicit confirmation.

draft · v0.1 · proposed

Apply a standing rule

Evaluate one situation and produce a logged authorized outcome or fallback.

draft · v0.1 · proposed

Maintain a personal glossary term

Preserve an owner-specific term, meaning, aliases, relationships, and lifecycle.

draft · v0.1 · proposed

Manage an organizing label

Maintain owner-confirmed labels used for organization and filtering.

draft · v0.1 · proposed

Detect a candidate glossary term

Identify a possible owner-relevant term without silently accepting it.

draft · v0.1 · proposed

Manage AI-assisted label vocabulary

Govern the non-authoritative vocabulary available for machine-applied cues.

draft · v0.1 · proposed

Apply an AI-assisted label

Add an attributable, removable soft-semantic cue to an eligible object.

draft · v0.1 · proposed

Apply contextual classification

Attach a Role or Mode context to an object with preserved lifecycle history.

draft · v0.1 · proposed

Maintain a deliberate knowledge note

Preserve owner-chosen knowledge as a versioned, linkable object.

draft · v0.1 · proposed

Maintain owner-authored profile knowledge

Preserve the owner's revisable self-description without treating it as observed inference.

draft · v0.1 · proposed

Maintain an evidence-backed profile assertion

Preserve a sourced, confidence-bearing and time-bounded assertion about a profile subject.

draft · v0.1 · proposed

Resolve a disputed profile assertion

Produce an owner-authorized, versioned disposition for contradictory profile knowledge.

draft · v0.1 · proposed

Apply an action-intent label

Attach a reusable handling intent without executing it.

draft · v0.1 · proposed

Invoke a governed action

Execute one declared operation under explicit authority and audit.

draft · v0.1 · proposed

Run a background operation

Carry long-running work to a durable terminal result without blocking the initiating surface.

draft · v0.1 · proposed

Configure execution policy

Govern which actions may run and when confirmation is required.

draft · v0.1 · proposed

Maintain an execution plan

Preserve a reviewable source design without confusing proposed work with execution.

draft · v0.1 · proposed

Activate an execution plan

Derive a project or routine from an eligible plan while preserving provenance.

draft · v0.1 · proposed

Manage a one-off project

Maintain a bounded structured effort through resolution.

draft · v0.1 · proposed

Maintain a reusable routine

Govern repeatable execution structure independently from its runs.

draft · v0.1 · proposed

Run a routine instance

Execute one attributable occurrence of a reusable routine.

draft · v0.1 · proposed

Manage an actionable task

Maintain one action through context, responsibility, state, and resolution.

draft · v0.1 · proposed

Organize a task collection

Group and order tasks without requiring project semantics.

draft · v0.1 · proposed

Schedule a contextual reminder

Re-surface an eligible object or message under a governed trigger.

draft · v0.1 · proposed

Prepare a context-rich Work Starter

Create a ready-to-start entry point over canonical work.

draft · v0.1 · proposed

Order ready work for the current context

Produce an explainable, revisable ordering of eligible work.

draft · v0.1 · proposed

Maintain a versioned workflow package

Define a distributable reusable process with an inspectable contract.

draft · v0.1 · proposed

Manage an installed workflow

Govern an owner's configured instance of a package version.

draft · v0.1 · proposed

Run a governed workflow

Execute one inspectable workflow instance without bypassing domain governance.

draft · v0.1 · proposed

Preserve an execution result

Register a produced outcome with provenance to its source execution.

draft · v0.1 · proposed

Record a canonical event

Append one idempotent, attributable occurrence to the event spine.

draft · v0.1 · proposed

Manage an owner attention item

Surface and resolve an owner-relevant information or response need.

draft · v0.1 · proposed

Publish a completed-work record

Create durable, evidence-linked completion records and owner-facing projections.

draft · v0.1 · proposed

Compile daily owner History

Produce the primary readable daily account from structured evidence.

draft · v0.1 · proposed

Aggregate an owner History period

Compose a higher-period summary from accepted lower-level History.

draft · v0.1 · proposed

Navigate time-indexed owner History

Traverse from broad periods to linked updates, events, and originals.

draft · v0.1 · proposed

Register a connected source

Govern a recurring external origin and its intake behavior.

draft · v0.1 · proposed

Manage an import session

Track one bounded intake event and its terminal outcome.

draft · v0.1 · proposed

Govern source promotion

Control staged movement from preservation to mapping, enrichment, and History eligibility.

draft · v0.1 · proposed

Screen an ingestion batch

Produce a bounded technical and safety review without granting authority.

draft · v0.1 · proposed

Assess ingestion readiness

Determine whether checklist conditions are closed for owner review.

draft · v0.1 · proposed

Authorize an ingestion batch

Grant or deny owner authority for one exact manifest and boundary.

draft · v0.1 · proposed

Govern an event-type registry

Maintain allowed event semantics, emitters, schemas, and lifecycle effects.

draft · v0.1 · proposed

Inspect an operational run record

Explain one run's state, authority, outputs, errors, and evidence without mutation.

draft · v0.1 · proposed

Compose a portable Core bundle

Package reconstructable owner state with manifests, provenance, and rebuild instructions.

draft · v0.1 · proposed

Compose a source-derived PIOS portability package

Package owner-controlled source material for validation and governed import without claiming prior Core state.

draft · v0.1 · proposed

Validate a portability package

Assess package safety, schema, integrity, semantics, ownership, and destination support.

draft · v0.1 · proposed

Restore Core state from a bundle

Hydrate accepted state while preserving identity, provenance, and explicit dispositions.

draft · v0.1 · proposed

Validate restored Core parity

Verify restored state, protections, provenance, boundaries, and selected behavior.

draft · v0.1 · proposed

Maintain a recoverable Core backup

Protect independent recovery copies and preserve their creation, integrity, custody, and retention status.

draft · v0.1 · proposed

Govern a Core cutover

Transfer canonical operating authority to a validated destination.

draft · v0.1 · proposed

Roll back a Core cutover

Restore the approved prior operating path after an aborted or reversed transition.

draft · v0.1 · proposed

Decommission a source Core

Retire the former source after parity acceptance and owner sign-off.

draft · v0.1 · proposed

Register an owner-data erasure request

Record bounded erasure intent and immediate exposure controls.

draft · v0.1 · proposed

Execute governed data erasure

Delete, expire, or render approved data inaccessible with verifiable dispositions.

draft · v0.1 · proposed

Reauthorize a destination connector

Establish new destination-scoped credentials, consent, and sync authority.

draft · v0.1 · proposed

Submit a governed content capture

Send selected content and source context through governed intake for an explicit receipt.

draft · v0.1 · proposed

Reconcile a pending capture

Resolve unaccepted capture work without silently discarding unique unconfirmed changes.

draft · v0.1 · proposed

Manage a guided onboarding journey

Maintain resumable staged progress through an explicit onboarding disposition.

draft · v0.1 · proposed

Maintain a personal AI identity

Govern the presentation identity through which a personal AI represents itself.

draft · v0.1 · proposed

Configure personal AI interaction preferences

Maintain how a personal AI communicates, explains, proposes, and collaborates.

draft · v0.1 · proposed

Prepare an initial owner environment

Produce a reviewable first-use configuration from owner input and explicit defaults.

draft · v0.1 · proposed

Authorize first entry

Issue and consume a bounded credential for first access to a prepared environment.

draft · v0.1 · proposed

Maintain an owner context Role

Govern a reusable life-context lens and its relationships over time.

draft · v0.1 · proposed

Maintain an owner operating Mode

Govern a reusable vocabulary for situational operating context.

draft · v0.1 · proposed

Set active personal context

Establish a temporary, provenance-bearing combination of current context dimensions.

draft · v0.1 · proposed

Configure a Role's outward intent

Declare a Role's private-only, privately shared, or public representation intent.

draft · v0.1 · proposed

Prepare a Role-scoped outward representation

Create a separate traceable representation for possible later outward use.

draft · v0.1 · proposed

Detect an observed pattern

Create a non-authoritative recurrence pattern from multiple governed evidence items.

draft · v0.1 · proposed

Evaluate an observed pattern lifecycle

Reassess pattern status, strength, evidence eligibility, and temporal state.

draft · v0.1 · proposed

Maintain an outcome goal

Preserve a durable longer-horizon desired outcome and its relationships.

draft · v0.1 · proposed

Maintain a directional objective

Preserve an active action-oriented direction that can guide planning.

draft · v0.1 · proposed

Maintain a success target

Define an observable condition describing what success would look like.

draft · v0.1 · proposed

Maintain a measurement definition

Define what and how to measure independently from observations and targets.

draft · v0.1 · proposed

Record a metric observation

Preserve one attributable time-scoped measured value or state.

draft · v0.1 · proposed

Share a governed outward representation

Deliver or activate a separate representation for a bounded audience under applicable policy and rights.

draft · v0.1 · proposed

Publish a governed outward representation

Make an eligible separate representation publicly reachable under governed policy.

draft · v0.1 · proposed

Maintain outward access policy

Establish, revise, suspend, or revoke the policy and grants governing access to an outward representation.

draft · v0.1 · proposed

Evaluate an outward access request

Decide whether one requester may open an outward representation under current policy.

draft · v0.1 · proposed

Govern outward discoverability

Control whether an outward representation may be found beyond a direct reference.

draft · v0.1 · proposed

Govern outward usage rights

Declare what human and machine recipients may do after access.

draft · v0.1 · proposed

Conduct a structured review

Evaluate a bounded scope and preserve findings, decisions, open loops, and carry-forward outcomes.

draft · v0.1 · proposed

Record a reflective assessment

Preserve attributable interpretation, uncertainty, residue, and recommendations over evidence.

draft · v0.1 · proposed

Maintain a human contact profile

Preserve an owner-held living profile for another person.

draft · v0.1 · proposed

Maintain a contact point

Preserve a typed communication or location coordinate for a person.

draft · v0.1 · proposed

Verify control of a contact point

Establish bounded evidence that a subject controls a declared contact point.

draft · v0.1 · proposed

Maintain personal relationship context

Preserve how the owner relates to another person or organization over time.

draft · v0.1 · proposed

Maintain an agent definition

Preserve the canonical owner-governed definition of a persistent software agent.

draft · v0.1 · proposed

Synchronize an agent definition to runtime

Reconcile a canonical definition version with declared runtime targets.

draft · v0.1 · proposed

Assign an agent runtime role

Bind an agent runtime to a bounded functional position and responsibility scope.

draft · v0.1 · proposed

Select an agent's outward identity

Govern identity presentation and AI disclosure for an outward action or action class.

draft · v0.1 · proposed

Conduct an owner-representative conversation

Let an authorized AI representative converse externally within bounded identity, audience, knowledge, and action scope.

draft · v0.1 · proposed

Maintain a collaboration channel

Organize durable human-agent discussions without becoming the canonical home of work or knowledge.

draft · v0.1 · proposed

Maintain a work-linked discussion thread

Preserve focused discussion and resolution context around a linked subject.

draft · v0.1 · proposed

Record an attributed collaboration message

Preserve one communication with true authorship, references, trust metadata, and trigger constraints.

draft · v0.1 · proposed

Manage a task-scoped execution workspace

Govern a bounded runtime work area while keeping temporary state separate from canonical outcomes.

draft · v0.1 · proposed

Maintain personal conversation history

Preserve continuous source-aware owner conversation history across sessions, contexts, interfaces, and admitted sources.

draft · v0.1 · proposed

Maintain an owner access path

Govern a route through which the owner accesses or steers a PIOS system.

draft · v0.1 · proposed

Maintain an authorized device-local content cache

Keep permitted local copies available and fresh without making the cache canonical.

draft · v0.1 · proposed

Maintain selected content for offline access

Retain explicitly chosen accepted content for a governed offline window.

draft · v0.1 · proposed

Manage a personal reading queue

Organize retained reading material through unread, saved, read, archived and removed states.

draft · v0.1 · proposed

Estimate operation resource cost

Forecast the resources or internal cost units likely required for a bounded operation.

draft · v0.1 · proposed

Reserve an operation resource budget

Hold a bounded internal resource allowance for one authorized operation.

draft · v0.1 · proposed

Reconcile operation resource usage

Compare actual resource use with estimates, reservations, budget, and completed scope.

draft · v0.1 · proposed

Discover a candidate information source

Identify reviewable source candidates without collecting or registering their content.

draft · v0.1 · proposed

Maintain owner implementation documentation

Preserve an owner-readable and editable account of the actual PIOS implementation.

draft · v0.1 · proposed

Assess knowledge-layer integrity

Evaluate knowledge for structural, evidence, semantic, lifecycle, and safety defects.

draft · v0.1 · proposed

Maintain a personal Meaning record

Preserve a contextual, revisable interpretation of why governed evidence matters to the owner.

draft · v0.1 · proposed

Maintain evidence-linked Learning

Preserve a revisable evidence-linked lesson that may help future reasoning.

draft · v0.1 · proposed

Govern a derived-representation lifecycle

Register and govern source-linked non-canonical representations through replacement, rebuild, revocation, and deletion.

draft · v0.1 · proposed

Promote a working artifact to canonical state

Move an accepted temporary output into its designated durable source-of-truth lifecycle.

Supporting documents