Skip to content
Skills Library
Data & analytics · Analysis

brain-knowledge

/brain-knowledge

Statements and corrections in Brain — the individual subject-predicate-object claims extracted from memories, their provenance, and their supersession history. Use when a stored fact is wrong, outdated or disputed, when the user asks where something came from or what changed and when, when a claim must be retracted rather than merely hidden, or when asserting a fact directly instead of waiting for extraction. Triggers on "that's wrong", "correct that", "where did you get that", "what changed", "since when", "retract", audit and provenance questions. Part of the Brain memory database; see brain

Repository
brain-kit ↗
Licence
Apache-2.0
Last repository change
06.11.2026

Assessment and security

90/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content25 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 06.11.2026, without references to retired models or APIs.

Licence

Apache-2.0, otwarta: można używać i zmieniać, z zachowaniem licencji i informacji o zmianach.

What it does and when to use it

Statements and corrections in Brain — the individual subject-predicate-object claims extracted from memories, their provenance, and their supersession history. Use when a stored fact is wrong, outdated or disputed, when the user asks where something came from or what changed and when, when a claim must be retracted rather than merely hidden, or when asserting a fact directly instead of waiting for extraction. Triggers on "that's wrong", "correct that", "where did you get that", "what changed", "since when", "retract", audit and provenance questions. Part of the Brain memory database; see brain

Triggers from the description

thatcorrect thatwhere did you get thatwhat changedsince whenretract
Original author description

Statements and corrections in Brain — the individual subject-predicate-object claims extracted from memories, their provenance, and their supersession history. Use when a stored fact is wrong, outdated or disputed, when the user asks where something came from or what changed and when, when a claim must be retracted rather than merely hidden, or when asserting a fact directly instead of waiting for extraction. Triggers on "that's wrong", "correct that", "where did you get that", "what changed", "since when", "retract", audit and provenance questions. Part of the Brain memory database; see brain

How it works

  1. Reading
  2. Correcting
  3. Asserting directly
  4. Supersession is a chain, not an overwrite
Original SKILL.md headings · 58 lines, 381 words
SKILL.md preview
---
name: brain-knowledge
description: Statements and corrections in Brain — the individual subject-predicate-object claims extracted from memories, their provenance, and their supersession history. Use when a stored fact is wrong, outdated or disputed, when the user asks where something came from or what changed and when, when a claim must be retracted rather than merely hidden, or when asserting a fact directly instead of waiting for extraction. Triggers on "that's wrong", "correct that", "where did you get that", "what changed", "since when", "retract", audit and provenance questions. Part of the Brain memory database; see brain-memory for the core store and recall loop.
---

# Brain knowledge

A memory is what was said. A **statement** is one claim drawn out of it —
subject, predicate, object — and statements are what `recall` actually ranks.
This skill is for working at that level: correcting a claim, tracing where it
came from, and seeing what it replaced.

## Reading

| | |
|---|---|
| `list_statements` | page through the facts extracted in this workspace |
| `get_statement` | one statement, with the memory it came from |
| `statement_history` | the full supersession chain, oldest to newest |
| `get_relation` | one typed edge between entities, with its endpoints |

`statement_history` is the one that answers "what did we think before". Reach
for it whenever the current answer looks older than it should — recall ranks
by **similarity, not recency**, so a stale claim can outrank the correction
that replaced it.

## Correcting

Three verbs, and the difference between them matters:

- **`supersede_statement`** — the claim changed. Both survive, linked, and
  the new one becomes current. **This is almost always the right one.**
- **`delete_statement`** — soft tombstone. The row survives for audit and
  stops being returned.
- **`retract_statement`** — the stronger form: the **content does not
  survive**. For something that should never have been recorded — a secret
  pasted by accident, a claim about a person that was false and harmful.

Use `retract_statement` only when the content itself is the problem.
Preferring it by default destroys the history that makes "what changed" a
question Brain can answer at all.

The same three exist for edges: `supersede_relation`, `delete_relation`.

## Asserting directly

`create_statement` and `create_relation` write a claim without waiting for
extraction to infer it. Use them when:

- the fact is structural and you already know it exactly, or
- extraction got it wrong and you are putting the right one in its place.

Prefer `remember` for anything a person said in prose. Extraction is what
turns one sentence into several linked claims; asserting by hand skips that
and gives you exactly what you typed and nothing more.

## Supersession is a chain, not an overwrite

Every correction keeps its predecessor. That is what makes provenance work —
and it is why the honest answer to "has this always been true?" is
`statement_history`, not a second `recall`.

Versions

The database has no other skills named brain-knowledge.

Database snapshot: 13.11.2026 · package version c413e463b422 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 90knowledge-correct
    Correct or withdraw specific saved Knowledge claims, explicit relationships, or pages at the user's request, preserving provenance and inspecting affected knowledge. Use for “that's wrong”, “retract that”, or a Knowledge chat reference; not automatic contradiction resolution or physical erasure.
  • 100brain-architect
    Architect and engineer of the user's claude-brain transcript database (~/claude-brain). Use to change the transcript schema or pipeline scripts, tune chunking/embeddings/ranking, migrate or rebuild the index, run a health check, clean up tags, or fix anything broken in the…
  • 100brain-architect
    Architect and engineer of the user's claude-brain transcript database (~/claude-brain). Use to change the transcript schema or pipeline scripts, tune chunking/embeddings/ranking, migrate or rebuild the index, run a health check, clean up tags, or fix anything broken in the…
  • 90brain-workspace
    Identity, sessions, spaces and query diagnostics for the Brain memory database. Use when a recall comes back empty and the workspace is in doubt, when the user asks which account or namespace is connected, when attributing writes to a session, when checking whether an LLM…
  • 90brain-mem
    Read and write the brain's memory correctly from inside a phase — the knowledge pages in your scope, a new or rewritten knowledge page filed in index.md under its tags, a proposal for system/, a project decision, the memory lint. Use whenever you read or change memory/…
  • 90knowledge-correct
    Correct or withdraw specific saved Knowledge claims, explicit relationships, or pages at the user's request, preserving provenance and inspecting affected knowledge. Use for “that's wrong”, “retract that”, or a Knowledge chat reference; not automatic contradiction resolution or physical erasure.

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.