agentleFS
NashEQifyNashEQify/forge/skills/knowledge_processor/SKILL.md

knowledge-processor

Skill2 starsChanged 3 days agoMITView on GitHub

Reconcile useful new information with scoped knowledge and retrieve relevant context before dependent work. Shared continuity infrastructure for all work; no software workflow or user repository is required for project knowledge.

We split this skill into the rules it gives an agent

5 rules. About 0 are common, 0 are distinctive, and 5 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Read and follow framework/work-principle.md before taking up work.

    line 17
  2. Very specific:

    When a user scope is configured or the project's rules require ongoing knowledge maintenance: read and follow …

    line 40
  3. Very specific:

    Before an answer, update or dependent continuation that needs stored context: read and follow skills/knowledge_processor/retrieve.md …

    line 48
  4. Very specific:

    Before admitting, writing, correcting, moving or removing a claim: read and follow skills/knowledge_processor/write.md (§2 admit, scope and …

    line 54
  5. Very specific:

    When writing history or the handoff, importing sources, or restructuring (splitting, merging or renaming) owners: read and follow …

    line 60

≈ “Common” is an estimate: at least 1 in 5 of the 308,232 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Read and follow framework/work-principle.md before taking up work. Resolve this reference against the framework root; retain the consumer CWD.

Knowledge Processor

Canonical semantic lifecycle for knowledge and continuity. Modes specialize this contract; boot, save and housekeeping invoke it. No database, hook, background worker, automatic import or development engine is implied.

Inputs and scope

Input: candidate information or retrieval need, active project, configured user scope if any, and mode process, user-signal or wrap-up. Resolve location and permissions through framework/user-scope.md. Project knowledge belongs to its verified project/topic owner; private person facts only to the resolved <user_repo>/context/user/. No user scope: project continuity still works; report unroutable person facts, never create a profile or private fallback in the project. A visiting agent retains its own user identity and follows the visited project's entry and authorization rules without adopting another person's profile.

Continuous maintenance

When a user scope is configured or the project's rules require ongoing knowledge maintenance: read and follow skills/knowledge_processor/continuous-maintenance.md at boot and after compaction (authority, same-cycle reconciliation, checkpoint before a topic switch, update-not-append, the FACTS line, blocked writes).

Executable lifecycle

1. Retrieve and establish a coherent baseline

Before an answer, update or dependent continuation that needs stored context: read and follow skills/knowledge_processor/retrieve.md (verified owner and baseline, bounded search and its stop, using what was retrieved).

2.–5. Admit, place, validate and publish

Before admitting, writing, correcting, moving or removing a claim: read and follow skills/knowledge_processor/write.md (§2 admit, scope and compare; §3 place and curate; §4 validity and provenance; §5 bounded impact and protected publication).

6. History, imports and maintenance

When writing history or the handoff, importing sources, or restructuring (splitting, merging or renaming) owners: read and follow skills/knowledge_processor/history-imports.md.

Modes and outcome

  • process: modes/process.md — project facts/impact; routes person facts separately.
  • user-signal: modes/user-signal.md — personal placement under the shared lifecycle.
  • wrap-up: modes/wrap-up.md — common continuity/history, selected domain contribution.

DONE: admitted deltas reconciled at authorized owners, routes/current views coherent, pending operation clear, and actual outcome verified. No-change PRESENT is success. FAIL: report unavailable scope, evidence, coherent snapshot or pending recovery; retain recoverable work in its scope, never claim completion on partial persistence. Close-phase retros supply their Patterns-Emerged as process input; do not re-extract raw sessions. Selected development closeout audit obligations remain domain-owned.

Incident handling: collect observations and locate/assess impact; defer transient current-state replacement until the incident's recovery/fix boundary. Preserve necessary incident checkpoints and approvals under local incident rules. Afterwards reconcile observed end state and relevant failed attempts. No live action is implicitly authorized by knowledge maintenance.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.