feat(adapters): add Kilo adapter (kilocode, an OpenCode fork)

CHANGELOG.md

@@ -9,6 +9,18 @@ breaking changes may land in a minor release.
 
 ### Added
 
+- **Kilo adapter (`kilo` profile, `kilo-http` kind).** Kilo (`kilocode`, an OpenCode fork)
+  is driven over its local HTTP/SSE server exactly like `opencode-http` — one headless
+  `kilo serve` per session, **no tmux window** — via a `KiloHttpAdapter` that extends the
+  OpenCode HTTP adapter. The fork differs from OpenCode in exactly two facts the adapter
+  overrides: it reads `KILO_*` env vars (`KILO_SERVER_PASSWORD`,
+  `KILO_CONFIG_CONTENT`, `KILO_DISABLE_EXTERNAL_SKILLS`) rather than `OPENCODE_*`, and its
+  server basic-auth username is `kilo`, not `opencode`. Everything else — the HTTP/SSE API
+  surface, SSE event frames, and usage schema — is identical to OpenCode. Needs the same
+  `[opencode]` extra for httpx; auth once globally with `kilo auth login`; skills live in
+  `.claude/skills/`; set `model` as `provider/model`. `env_fault_patterns` are intentionally
+  unseeded pending a captured kilo outage line.
+
 - **`repo_root` in run `state.json`** (#716). A run records the git root its code work happens in,
   so an out-of-process reader — `bmad-loop resolve`'s re-arm — uses the tree the run measured
   instead of re-deriving one. A `state.json` written before the field existed degrades to the

README.md

@@ -436,7 +436,7 @@ skill = "bmad-dev-auto"    # the only supported value — the generic upstream d
                            # No settings-schema entry: edit it here, not in the TUI editor.
 
 [adapter]
-name = "claude"            # CLI profile: claude | codex | gemini | copilot | antigravity | opencode-http (alias: opencode) | custom
+name = "claude"            # CLI profile: claude | codex | gemini | copilot | antigravity | opencode-http (alias: opencode) | kilo | custom
 model = ""                 # empty = CLI default (opencode-http wants "provider/model")
 cleanup_session_on_finish = true  # kill the run's tmux session when it finishes (false keeps it for inspection)
 # extra_args replaces the profile's default bypass flags when set:
@@ -646,6 +646,7 @@ One generic driver (`adapters/generic.py`) runs any coding CLI that fits the inj
 | `copilot`     | supported, E2E-verified                  | GitHub Copilot **CLI** (the `copilot` binary, GA ≥ 2026-02) — _not_ the VS Code extension. Launches with `-i` to stay interactive; turn-end is `agentStop` (per response turn); `--allow-all-tools` for unattended runs. `copilot-events` usage parser reads token totals from the trailing `session.shutdown` line, so the profile waits a short grace (`usage_grace_s = 8`) before tallying. **Pin a capable model** (see below).                                                                                                                                                                                                                                                                                                                                                                                                                                                        |
 | `antigravity` | experimental — `isolation = "none"` only | Google **Antigravity CLI** (`agy` ≥ 1.1.3). Launches with `-i` to stay interactive; `Stop` is the turn-end event (agy has no SessionStart/SessionEnd hook). Skills and hooks live in `.agents/` (flat `Stop` handler in `.agents/hooks.json`, keyed by hook-group name). Hook payloads are protojson/camelCase. **Trust is exact-path**: `agy` blocks on a "trust this folder" dialog for any workspace not listed verbatim in `settings.json` `trustedWorkspaces`, and `--dangerously-skip-permissions` does not bypass it — so `isolation = "worktree"` hangs ([#169](https://github.com/bmad-code-org/bmad-loop/issues/169)). `usage_parser = "none"` is permanent, not pending: agy's transcript carries no usage data (tokens live only in an internal SQLite/protobuf store), so runs work but token columns stay empty. Verify against your build with `probe-adapter antigravity`. |
 | `opencode`    | supported, E2E-verified                  | **OpenCode** ≥ 1.18 (profile `opencode-http`), driven over HTTP/SSE — one headless `opencode serve` per session, **no tmux window**. Needs the extra: `pip install 'bmad-loop[opencode]'`. Auth once globally with `opencode auth login`; skills live in `.claude/skills/`; set `model` as `provider/model` (e.g. `anthropic/claude-haiku-4-5`). Watch sessions via `run_dir/logs/<task_id>.log` or the TUI Log tab; `resolve` is `--no-interactive` only; the Unity plugin's window guards are unsupported here.                                                                                                                                                                                                                                                                                                                                                                          |
+| `kilo`        | supported                                | **Kilo** (`kilocode`, an OpenCode fork, profile `kilo`), driven over HTTP/SSE exactly like OpenCode via the `kilo-http` adapter — **no tmux window**. Needs the same extra: `pip install 'bmad-loop[opencode]'`. Auth once globally with `kilo auth login`; skills live in `.claude/skills/`; set `model` as `provider/model`. The fork reads `KILO_*` env vars and authenticates with the basic-auth username `kilo` (not `opencode`), which is all the `kilo-http` adapter changes from the OpenCode family.                                                                                                                                                                                                                                                                                                                                                                             |
 
 **Copilot — pin a capable model:** Copilot's free default (GPT-5 mini) is unreliable for the multi-step dev/review skills — it silently skips steps mid-workflow and fails the story. Set a capable model in policy, e.g. `[adapter] model = "claude-sonnet-4-6"` (passed through as `--model`), for end-to-end reliability. Because Copilot fires `agentStop` per response turn, a thorough multi-turn review needs more than one nudge to finish; the profile ships `stop_without_result_nudges = 5`, and you can tune it per stage (e.g. `[adapter.review] stop_without_result_nudges = …`). Both knobs are editable in the settings TUI under `[adapter]`.