remember

name	remember
type	skill
category	instruction
description	Unified memory skill: immediate mode (/remember) persists knowledge via PKB MCP; maintenance mode (/sleep, GHA cron) runs periodic consolidation — transcript mining, knowledge synthesis, data quality, brain sync.
triggers	["remember this","save to memory","store knowledge","sleep cycle","consolidation","brain maintenance"]
modifies_files	true
needs_task	false
mode	execution
domain	["operations"]
allowed-tools	mcp__pkb__create,mcp__pkb__create_memory,mcp__pkb__append,mcp__pkb__get_document,mcp__pkb__search,mcp__pkb__update_task,Bash,Read,Write,Grep,Glob,Edit,Skill,mcp__pkb__pkb_orphans,mcp__pkb__list_tasks,mcp__pkb__graph_stats,mcp__pkb__get_stats,mcp__pkb__get_network_metrics,mcp__pkb__get_task,mcp__pkb__task_search,mcp__pkb__pkb_context,mcp__pkb__batch_reparent,mcp__pkb__find_duplicates,mcp__pkb__batch_merge,mcp__pkb__merge_node,mcp__pkb__complete_task,mcp__pkb__batch_reclassify,mcp__pkb__batch_archive,mcp__pkb__batch_update,mcp__pkb__create_task,mcp__pkb__update_memory,mcp__omcp__messages_search,mcp__omcp__messages_query,mcp__omcp__calendar_list_events
owner	pauli
version	4.0.0

Memory Skill

Provides domain instructions for knowledge capture, persistence, and maintenance in the Personal Knowledge Base (PKB) ($ACA_DATA).

Immediate Mode (/remember)

Invariants

• Do not use Write or Edit on files under $ACA_DATA/. Use the PKB MCP toolset.
• Use mcp__pkb__search for semantic discovery. Do not use Glob or Grep on $ACA_DATA/.

Hard Rules

• mcp__pkb__create: Create a document (projects, contexts, meeting notes).
• mcp__pkb__create_memory: Add a memory/note.
• mcp__pkb__append: Extend an existing document.
• mcp__pkb__get_document: Read a document.
• mcp__pkb__search: Search before creating new content — and before asserting a fact. Don't assume; look it up.

Curation

• Capture episodic records accurately in the moment.
• Curate and synthesize knowledge inline as decisions or patterns emerge; do not defer all synthesis to the /sleep loop.

Storage Hierarchy

Content	Write Via	Target Location
Epics/projects	mcp__pkb__create(type="epic"|"project")	projects/
Tasks/issues	gh issue create (GitHub primary)	Synced automatically
Durable knowledge	mcp__pkb__create or mcp__pkb__create_memory	knowledge/, context/
Session findings	mcp__pkb__update_task on parent task	Task body
Axioms / framework rules	Invoke /framework skill — do NOT write to brain/context/	skills/ via framework workflow

File Locations

Content	Location	Notes
Project metadata	projects/<name>.md	Hub files. Hub-doc type is epic.
Project details	projects/<name>/	Subdirectory.
Goals	goals/	Strategic objectives.
Context (about user)	context/	User preferences, history.
Sessions/daily	sessions/	Daily notes only, type: daily-note.
General knowledge	knowledge/<topic>/	Durable facts (non-user).
Meeting/call notes	knowledge/<topic>/ or projects/	type: meeting-note.
Maps of Content (MOCs)	knowledge/ or topic subdirs	Navigational hubs, type: moc.

Episodic Classification

• Operational episodic (agent activity, debugging logs, experimental steps) -> Update parent task body using tasks MCP tools. Do not create new documents in $ACA_DATA/.
• Durable episodic (meeting notes, user daily notes) -> Save to $ACA_DATA/ with correct type frontmatter.

Canonical Topic Notes (Enduring Memory)

Maintain exactly one canonical note per first-class topic (tools, projects, skills, agents, concepts).

• Update existing topic notes using mcp__pkb__append in the relevant section.
• Create new canonical notes with a standard section scaffold (e.g., Overview, Usage, Known Issues) only if the topic does not exist.
• Avoid creating one-off observation files.

Reconciliation

During updates, search for duplicate or peer notes. Merge unique content, retire weaker versions, and update wikilinks. For how to retire — delete the superseded file and trust git history; do not leave superseded_by: pointers — follow the single canonical rule in [[consolidation-procedure]] Step 3.

Workflow

1. Search: Search PKB first: mcp__pkb__search(query="topic").
2. Evaluate: If a canonical note exists, append to it. Do not create a separate file.
3. Write: Call the appropriate PKB MCP tool:

mcp__pkb__create(
  title="Descriptive Title",
  body="Content with [[wikilinks]].",
  type="note" | "project" | "epic" | "knowledge" | "moc" | "meeting-note",
  tags=["tag1", "tag2"]
)

Or for lightweight facts:

mcp__pkb__create_memory(
  title="descriptive title",
  body="content",
  tags=["tag1"]
)

Graph Integration

Links and References

• Every note must contain at least one [[wikilink]] to a parent concept, project, goal, or entity in prose.
• Link all proper nouns.
• Provide explicit markdown links for external issues and PRs (e.g., [org/repo#123](url)).
• List semantic relationships in a ## Relationships section:

  ## Relationships
  
  - [related] [[task-id]] — description
  - [upstream-bug] [org/repo#123](url)
  

• Wikilink Conventions: Add wikilinks only in prose. Never put them inside code blocks, inline code, or technical tables. Do not create a standalone "See Also" section.

Abstraction Level

For framework tasks, capture generalizable patterns (e.g., "Configuration should be set once at initialization") rather than local implementation details.

Observation Notation

Factual claims in episodic notes must be formatted as Obsidian callouts:

> [!observation] Factual statement
> Source: [[source-note]]
> Confidence: <numeric value 0.0 to 1.0>

• Numeric confidence mappings (include numeric value in frontmatter):
  • established (>= 0.8): verified, multiple sources.
  • provisional (0.4 - 0.79): limited evidence.
  • speculative (< 0.4): inference.
• If a new observation contradicts an existing one, record both with sources and flag the contradiction for the user. Do not overwrite silently.

Provenance (Required for type: knowledge)

Include these frontmatter fields:

• sources: (YAML list)
• synthesized: (ISO date)
• last_reviewed: (ISO date)
• confidence: (numeric 0.0 - 1.0)
• maturity: seedling | budding | evergreen

Maps of Content (MOCs)

Use MOCs to index and curate links to related notes when a topic area grows to 5+ notes.

MOC Frontmatter

title: "MOC: Topic Name"
type: moc
tags: [moc, topic-area]
created: YYYY-MM-DD
last_reviewed: YYYY-MM-DD

Background Capture

To capture memories asynchronously:

Task(
  subagent_type="general-purpose", model="haiku",
  run_in_background=true,
  description="Remember: [summary]",
  prompt="Invoke Skill(skill='remember') to persist: [content]"
)

Output Expectations

State the PKB write details directly and concisely:

• Tool used
• Title of note
• ID / permalink (returned by PKB) Do not expose raw local filesystem paths.

Maintenance Mode (/sleep)

Runs periodic offline consolidation on a GHA workflow or manual trigger.

Rules

• Never fabricate assertions without provenance.
• Do not edit daily notes, meeting notes, or task bodies (except adding consolidated: YYYY-MM-DD in frontmatter).
• Record and preserve contradictions.

Mode Detection

First match determines mode:

1. Environment variable SLEEP_MODE overrides default (short-loop | full-session).
2. LOOP_INTERVAL_MINUTES <= 30 -> short-loop.
3. If prior cycles ran in this session -> short-loop.
4. Default -> full-session.

Maintenance Phases

See [[references/maintenance-phases]] for detailed sub-agent rules, pacing, and exit protocols.

Phase	Name
0	Graph Health Baseline
1	Session Backfill
2	Transcript Mining
3	Episode Replay
4	Knowledge Consolidation
5	Index Refresh
6	Data Quality Reconciliation
7	Staleness Sweep
8	Refile Processing
9	Graph Maintenance
10	Consolidation Self-Check
11	Brain Sync

CI Environment Constraints

When running on GitHub Actions, no PKB MCP server is available. Direct file tool operations (Bash, Glob, Grep, Read, Write, Edit) on markdown files are allowed. Do not edit files outside $AOPS_SESSIONS/ (except git commits/push).