Skip to content
Forgeplan

forgeplan fpf

forgeplan fpf is the parent command for the First Principles Framework (FPF) Knowledge Base — a 204-section reasoning corpus that powers Forgeplan’s ADI cycle, trust calculus, and explore/investigate/exploit rule engine.

FPF is the theoretical backbone behind forgeplan reason, R_eff scoring, and methodology integrity checks. The fpf subcommands let you load the spec, query it semantically, inspect active rules, and check how those rules apply to specific artifacts.

When to use

Section titled “When to use”
  • Once per workspace — run fpf ingest after forgeplan init so the KB is available locally.
  • During reasoning — fpf search "trust calculus" to pull first-principles context while shaping a PRD or ADR.
  • During sprint planning — fpf dashboard to see bounded contexts, quality scores, and explore-vs-exploit recommendations.
  • During validation — fpf check PRD-XXX to see which FPF rules fire on an artifact and what action they suggest.
  • For onboarding — fpf list + fpf section B.3 to read the spec directly from the CLI.

When NOT to use

Section titled “When NOT to use”
  • For general artifact CRUD — use forgeplan new, validate, review, activate instead.
  • For project-wide health — use forgeplan health, not fpf dashboard (the two are complementary, not interchangeable).
  • For ADI reasoning runs — use forgeplan reason --fpf, which internally consults the KB; direct fpf search is for humans inspecting raw content.

Usage

Section titled “Usage”
forgeplan fpf <COMMAND>

Options

Section titled “Options”
  -h, --help     Print help
  -V, --version  Print version

Subcommands

Section titled “Subcommands”
  dashboard  Show FPF dashboard — bounded contexts, quality scores, explore-exploit actions
  ingest     Ingest FPF spec into knowledge base
  search     Search FPF knowledge base
  section    Show a specific FPF section
  list       List all FPF sections
  status     Show FPF knowledge base status — source, ingested count, staleness
  rules      List active FPF rules grouped by action (EXPLORE/INVESTIGATE/EXPLOIT)
  check      Check which FPF rules match a given artifact
  help       Print this message or the help of the given subcommand(s)

Examples

Section titled “Examples”
Terminal window
# One-time setup after forgeplan init
forgeplan fpf ingest


# Explore the KB
forgeplan fpf status
forgeplan fpf list
forgeplan fpf section B.3


# Pull first-principles context into your reasoning
forgeplan fpf search "trust calculus"
forgeplan fpf search "bounded context"


# Apply FPF rules to the project and specific artifacts
forgeplan fpf dashboard
forgeplan fpf rules
forgeplan fpf check PRD-019

How it fits

Section titled “How it fits”

FPF sits at the reasoning layer of the Forgeplan pipeline:

Shape → Validate → ADI (FPF KB) → Code → Evidence → Activate
  • PRD-041 wires FPF rules into the route/validate stages.
  • PRD-042 adds BGE-M3 vector search across the 204 sections (the same pipeline as the artifact search — BM25 + semantic fusion).
  • PRD-043 enforces methodology integrity: an artifact that violates bounded-context or trust-calculus rules is flagged before activation.

The KB is stored in LanceDB under .forgeplan/lance/ (derived, gitignored). Raw sections live in the Forgeplan repo and are embedded on ingest.

See also

Section titled “See also”