This repo uses Clauderizer for durable, cross-session working memory. Markdown
under docs/ is the source of truth; an MCP server exposes it as live tools.
Start of every session — load memory first. If a [Clauderizer] status digest
appeared at the top of this session, your host has a session hook and memory is
already loaded. If no digest appeared, call cz_status now, before anything
else — many hosts have no session hook, and cz_status is your durable project
memory (where the gameplan stands, the baseline, open items). Either way, to
begin or continue a phase, call cz_next_phase_context then cz_preflight.
On a host with no session hook NOTHING is injected automatically: the
disciplines marked instructions-floor in docs/ENFORCEMENT.md are carried by
this very text plus a one-time status note on your first tool result — keep
this block in your agent-instructions file.
If the cz_* tools themselves are absent (the call fails, not just a missing
digest), the wiring is broken, not the memory: run clauderize doctor in a shell,
use clauderize status for the digest, and make tracked writes with clauderize ops <file.json|-> — a JSON batch [{"op": "cz_add_lesson", "args": {...}}, ...]
against the exact tool names and schemas — until the wiring is repaired. Every
cz_* tool is reachable that way.
Key tools (MCP server clauderizer):
cz_status/cz_next_phase_context— where things stand; the next phase bundle.cz_graph_query— look up an entity and its dependents/dependencies.cz_preflight— run the pre-flight checks (tests/build via the host profile).cz_cascade/cz_resolve_cascade— after a tracked edit, walk dependents, then record the verdicts (never hand-edit the report).cz_write_handoff— assemble the next cumulative phase handoff (your notes outside its marker block survive regeneration).cz_add_decision/cz_add_invariant/cz_add_lesson/cz_add_correction/cz_upsert_entity/cz_transition_status— structured, graph-aware writes.cz_add_output/cz_add_phase_summary— the Ending Protocol writes: record a phase's concrete produced values and its completion summary.cz_consolidate_lessons/cz_promote_lesson/cz_obsolete_lesson— keep memory compact: synthesize overlap, promote the enduring todocs/LESSONS.md, mark the stale.cz_add_dream— after each substantive exchange, leave a 2–4 sentence dream note (kind: friction | gap | surprise | correction | drift | win; no PII, use repo-relative paths) in the local journal — the dreamer distills accumulated notes into next-session proposals.cz_add_open_item/cz_resolve_open_item/cz_set_exit_criteria/cz_check_exit_criterion/cz_analyze— the discipline gates (D-015/D-016, always advisory & judgment-based — INVARIANT-05): track open items and machine-checkable exit criteria (surfaced when a phase completes), and surface the decisions/invariants most relevant to a new decision for contradiction-judgment.
Rules: never hand-edit frontmatter or append to tracked logs directly — use the
cz_* tools (or clauderize ops when MCP is unavailable) so the graph stays
consistent. The procedure spec is at docs/gameplans/GAMEPLAN-PROCEDURE.md.