Concepts
Explanations of ideas, principles, and mental models. These documents answer “what is X and why does it work this way?”
If you need a step-by-step guide, see guides/. If you need a normative spec, see contracts/.
Contents
| Document | What it explains |
|---|---|
| overview.md | What AletheIA is and how the repo is structured |
| operating-overlay.md | The overlay layer: what it is, what belongs in it, what doesn’t |
| architecture.md | Core contracts and canonical surfaces |
| ecosystem-map.md | Where AletheIA and Adaptive Skills fit in the broader agentic tooling ecosystem |
| ecosystem-territory-map.md | Non-normative north-star view of the internal territory being built across both repositories |
| canonical-vocabulary.md | Stable shared language: Work Slice, Restart Package, Operational Boundary, Handoff, etc. |
| governance.md | What is mandatory, forbidden, and requires approval |
| quality.md | Quality baseline — what the framework must prove |
| enforcement-boundaries.md | Behavioral vs technical enforcement |
| durable-decisions.md | How important decisions survive sessions |
| work-slice-pattern.md | The primary operational unit |
| work-item-pattern.md | External coordination abstraction |
| agent-handoffs.md | Cross-agent and cross-session transitions |
| handoff-capture-pattern.md | How to capture handoff signals from completed work |
| slice-finalization-and-restart.md | → see guides/ |
| agent-harness-contract.md | The declared per-task operating envelope (autonomy, tools, gates, sensors, rollback) |
| autonomy-levels.md | The four canonical authority levels and how external five-level drafts map onto them |
| tool-risk-taxonomy.md | Coarse risk classes for skill declaration, mapped to the authoritative permission matrix |
| agent-computer-interface.md | How to design tools agents consume (ACI), complementing the permission matrix |
| context-rot-controls.md | Signals of long-session degradation and the minimal controls + checkpoint |
| execution-pattern-governance.md | How AletheIA chooses the execution topology before task execution |
| execution-pattern-library.md | The ten canonical execution patterns and selection rules |
| execution-vehicle-selection.md | Vehicle vs pattern and proportionality before autonomy |
| comprehension-debt.md | When generated material outpaces human understanding, distinct from context rot |
| self-application.md | How AletheIA governs its own evolution |
| structured-risk-inference.md | Experimental risk inference layer |
| project-extension-pattern.md | How projects extend the overlay without distorting it |
| project-local-constitution-context.md | Project-local constitution layer |
| local-trust-boundary-posture.md | Trust-boundary posture as project-extension concern |
| iterative-maintenance-governance.md | Governance across maintenance rounds |
| progressive-policy-signals.md | From telemetry to reviewable policy signals |
| visual-operations-layer.md | Read-only projection that makes governed Work Slices operationally visible |
| mission-control-cockpit.md | Human-facing Mission Control views, empty states, alerts, and evidence posture |
| context-graph-integration.md | How a code graph fits into the operating model |
| bootstrap-principles.md | Alpha 7 bootstrap posture |
| distribution-presets-adapters.md | Alpha 6 distribution direction |
| adapter-taxonomy.md | Taxonomy for AletheIA adapters |
| preset-taxonomy.md | Taxonomy for AletheIA presets |
| domain-governance-packs.md | Future domain-specific governance packs |
| ai-agent-security-prompt-injection.md | Concept source for AI agent security; S28 pack lives in domain-governance-packs/ |
| web-app-security-trust-boundaries.md | Concept source for web/API trust boundaries; S29 pack lives in domain-governance-packs/ |
| knowledge-governance-layer.md | Layer that governs user-provided knowledge bases used by agents and skills |
| user-provided-knowledge.md | How users register knowledge bases without dumping documents into prompts |
| framework-capsules.md | Operational summary of a framework or large source; the default unit a skill consumes |
| knowledge-resolver.md | Logical role that selects which knowledge sources enter task context |
| governed-skill-optimization.md | Why skill optimization is governed: evidence vs. authority, and where the boundary sits |