Skip to content

graff — a fast agentic coding harness in Zig: multi-provider, MCP, workflows, DGM evolution loop, TS/Python SDKs

Python
3.8 KB
Raw Blame

yxlyx leftovers (279 continuation)

Historical status at the v0.0.282 cut: what existed in this repo versus what was parked. This is not a current issue inventory; the embedded TUI and its simulator have since moved out of this repository. No new product mode.

Landed on v0.0.282: #674 atomic paste spans, #693 Codex WS type:error, #694 HTTP client generations (ADR 0048), #697 clone-on-write session branches (ADR 0049; #698 was a yxlyx reopen of the same head and is closed as duplicate), #679 MCP TLS reuse (ADR 0050), #695 omit an empty tool catalog, #677 /teleport + /snapshot gc (ADR 0051), #675 DeepSeek thinking-off / lean 15s bash (ADR 0052–0055). Still parked below.

Issue In this repo today This cut
#321 /doctor Goal/todo slice plus the live job table and RunBudget: JOB_STATE, JOB_SHARES_ROOT_PGID, BUDGET_STATE, BUDGET_NEAR_LIMIT, BUDGET_EXCEEDED. TUI /doctor uses the same report (no chrome-only "ok" line). Partial. Parked: durable session-lease, listener, and orphan-descendant findings (DUPLICATE_WORKTREE_OWNER, STALE_SESSION_LEASE, ORPHANED_OWNED_JOB, LISTENER_AFTER_CLEANUP). Those need a registry that does not exist; doctor still refuses to fake them.
#217 governed run budget src/run_budget.zig: invocation-wide --max-model-calls (default unlimited), concurrency 8, depth 1. Charged across root / children / title / judges / compaction. Exhaustion is a structured stop, not a model plea. --max-tool-calls is still root-only. Partial. Parked: finite defaults for /goal//loop, wall-time, token, and USD ceilings, descendant tool accounting, slash widening-with-approval.
#220 protected verifiers --eval / --until scores a command the agent can invoke. ADR 0008 is the eval-suite rule, not a /goal acceptance contract. Parked. Independent verifier stages and hidden tests are a new controller, not a docs tweak.
#306 review runaway Same budget knobs as #217. /review is one isolated pass (catalog). No review-specific cycle cap or "confirm before implementing" gate. Parked on the #217 remainder.
#283 cube NDJSON buffer graff serve already emits NDJSON as events happen. The failure is Daytona preview buffering the response and iOS URLSession idling out. graff cube / src/cube.zig is the sandbox+serve CLI. Closed, not planned (the in-repo iOS app was removed).
#199 idle localhost servers No graff processes / idle supervisor. Jobs exist for tool-started bash; they do not pause forgotten next dev trees. PR #200 closed as abandoned vs tip. Parked. Do not start a process-supervisor product on this branch.
#106 representative Ultracode sims Unit suite + TUI/sim.zig Term + tier-2 harness cases. Not a multi-turn human Ultracode drama. Parked (eval/tier-2 work, not 279 continuation).

#218 / #219 (siblings under #216 governed runs) follow #217/#220: do not open a second budget product here.

Cheap truth

The leftover that already lives as product is /doctor: goal/todo plus the live in-process job table and RunBudget (JOB_STATE, JOB_SHARES_ROOT_PGID, BUDGET_STATE, BUDGET_NEAR_LIMIT, BUDGET_EXCEEDED). Line REPL and TUI share that report. Everything else is another surface (Daytona, iOS, a durable lease/listener/orphan registry) or a new mode (idle-server supervisor, protected verifier, review cycle cap).