| name | aethermind-continuity |
| description | Use when a Hermes session needs to create, read, validate, or reorient from project-local AetherMind continuity layers. |
AetherMind Continuity
AetherMind stores compact, append-only .aem records beside a project. Use it
for load-bearing continuity: decisions, corrections, uncertainty, pressure,
rationale, and evidence pointers that future agents need before touching the
same project.
Before Writing
- Identify the real project root.
- Prefer compact continuity signals over transcripts, logs, or task summaries.
- Do not write secrets, credentials, private paths, raw prompts, transcripts,
or high-volume logs into
.aem layers.
Common Tools
aethermind_init_store creates .aethermind/layers.aem and
.aethermind/texture.aem.
aethermind_write_layer appends a structured continuity layer.
aethermind_read_layers searches stored layers.
aethermind_reorient builds a task-relevant continuity bundle.
aethermind_evaluate_store validates schema, privacy, density, and continuity
properties.
Layer Guidance
Use layer types intentionally:
load-bearing for decisions or constraints that future agents must preserve.
correction for mistakes, changed assumptions, or repeated failure modes.
friction for pressure that shaped the work.
uncertainty for unresolved facts that change the next step.
discovery for evidence-backed findings.
fork for alternate paths or abandoned branches.
Keep each layer short enough to scan quickly.