mirror of
https://github.com/AgentSeal/codeburn.git
synced 2026-08-23 15:34:19 +00:00
c7c3a878d8
flatSlice returned strings within the bound unchanged, but provider adapters pre-truncate with .slice(0, 500) before the cache site, so those views still pinned their parent buffers. Always flatten; the round-trip is ~150ns per turn. Use utf16le so lone surrogates survive the copy. Cache the canonical-path Promise instead of the resolved value so calls in one Promise.all batch share a single walk. Document the one-time kiro re-parse and worktree regrouping.
68 lines
6.6 KiB
Markdown
68 lines
6.6 KiB
Markdown
# Kiro
|
||
|
||
Kiro IDE chat history.
|
||
|
||
- **Source:** `src/providers/kiro.ts`
|
||
- **Loading:** eager (`src/providers/index.ts:7`)
|
||
- **Test:** `tests/providers/kiro.test.ts`
|
||
|
||
## Where it reads from
|
||
|
||
VS Code-style globalStorage at `kiro.kiroagent`:
|
||
|
||
| Platform | Path |
|
||
|---|---|
|
||
| macOS | `~/Library/Application Support/Kiro/User/globalStorage/kiro.kiroagent` |
|
||
| Windows | `%APPDATA%/Kiro/User/globalStorage/kiro.kiroagent` |
|
||
| Linux | `~/.config/Kiro/User/globalStorage/kiro.kiroagent` and `~/.kiro-server/data/User/globalStorage/kiro.kiroagent` (remote dev boxes / Cloud Desktop) |
|
||
|
||
On Linux both the local and remote-server globalStorage paths are scanned when present.
|
||
|
||
Two more stores live outside globalStorage, under `~/.kiro/sessions` (honors `KIRO_HOME`):
|
||
|
||
- **CLI store:** `~/.kiro/sessions/cli/<id>.jsonl` (+ companion `<id>.json`)
|
||
- **IDE v2 store:** `~/.kiro/sessions/<workspace-hash>/sess_<id>/` (see below)
|
||
|
||
Sessions in globalStorage are under hash-named workspace subdirectories. Discovery keeps backward compatibility with legacy `.chat` files and also scans the post-February 2026 extensionless format:
|
||
|
||
- `<workspace-hash>/<execution-id>.chat` legacy session files
|
||
- `<workspace-hash>/<session-hash>` extensionless session index files
|
||
- `<workspace-hash>/<session-hash>/<execution-hash>` extensionless execution files inside session directories
|
||
- `workspace-sessions/<base64url(path)>/<uuid>.json` v1 session records, a sibling of the workspace-hash directories at the `kiro.kiroagent` root (conversation spine; execution-backed stubs are skipped to avoid double-counting the execution files)
|
||
|
||
## Storage format
|
||
|
||
Kiro has several known formats across generations:
|
||
|
||
- **Legacy `.chat`** files with `{ chat, metadata, executionId }`
|
||
- **Modern extensionless execution files** with identifiers/timestamps at the top level plus conversation fields such as `messages`, `conversation`, `chat`, `transcript`, `entries`, `events`, `context.messages`, or direct prompt/response fields. In v1 these are the *content* half; the paired `workspace-sessions/<uuid>.json` record is the *spine* (linked by `executionId` ⇄ `chatSessionId`).
|
||
- **CLI** JSONL: one `{ version, kind, data }` event per line + a companion `<id>.json` with `session_state` metering (`metering_usage` gives real per-turn cost in credits).
|
||
- **IDE v2** (`sess_<id>/`): a self-contained, event-sourced session directory — `session.json` (metadata incl. the real `modelId`, `title`, `workspacePaths`) + `messages.jsonl` (append-only `{ id, timestamp, payload:{ type, ... } }` events) + `publish.cursor` + `snapshots/`. Assistant content is inline (no separate execution files). Usage is billed in **credits** via `usage_summary.promptTurnSummaries` (no token counts): cost comes from credits at the public overage rate ($0.04/credit, `USD_PER_KIRO_CREDIT`), while token counts are estimated from transcript text.
|
||
|
||
Session index files with `{ executions: [...] }` are discovered but skipped during parsing because they do not contain conversation content.
|
||
|
||
## Caching
|
||
|
||
None.
|
||
|
||
## Deduplication
|
||
|
||
Modern files deduplicate per session/execution pair. Legacy `.chat` files deduplicate per workflow/execution pair. CLI turns key on `kiro-cli:<sessionId>:<turnIndex>`. v2 turns key on `kiro-v2:<sessionId>:<executionId>`.
|
||
|
||
The stores are disjoint (v2 sessions use `sess_`-prefixed IDs in a separate directory), so no session is counted twice today. Migration from v1 workspace-sessions to v2 is currently detect-only; if it begins moving records, add a guard preferring v2 when the same logical session appears in both.
|
||
|
||
## Quirks
|
||
|
||
- **Workspace hash resolution** is non-trivial. The parser tries `workspace.json` first; if that fails, it base64-decodes the directory name to recover the workspace path.
|
||
- **Model ID normalization.** Kiro stores models like `claude-1.2`; the parser rewrites the dot to a hyphen so they match `claude-1-2` in the pricing snapshot. Add new versions here when Kiro ships them.
|
||
- **Tool name extraction accepts text and structured calls.** Kiro can embed tool calls inside message text as `<tool_use><name>...</name>` or expose structured `toolCalls` / `tool_calls` / `tools` entries.
|
||
- Token counts are estimated via char count (`CHARS_PER_TOKEN = 4`).
|
||
- **Credits are the cost source; tokens stay estimated.** Kiro bills in credits ($20/mo for 1,000; overage $0.04/credit). CLI (`metering_usage`), v1 executions (`usageSummary[].usage`), and v2 (`usage_summary.promptTurnSummaries[].usage`) turns record real credits, converted to USD at `USD_PER_KIRO_CREDIT = 0.04` (the public overage rate — the same never-understate approach as Codebuff). Turns without credit data fall back to token-estimated cost (`costIsEstimated: true`); legacy `.chat` and workspace-session records carry no usage data, so they are always token-estimated. Note: an earlier CLI implementation summed credit values directly as dollars, overstating cost 25×. Token *counts* remain char-estimated everywhere (input undercounts: only visible transcript text is seen, not the full resent context; v2's `session_metadata.contextUsage.usagePercentage` × context window is a better input proxy if ever needed). v2 does keep the real `modelId`, so unlike the v1 execution-file path it is not mislabeled `kiro-auto`.
|
||
- **Cost is frozen at parse time.** Kiro is on the `costUSD` pass-through allowlist in `providerCallToCachedCall` (alongside mistral-vibe, devin, hermes, …), so its credit-based cost survives the session cache instead of being re-priced from estimated tokens — token re-pricing understated/overstated real kiro spend by up to 16× per model. The tradeoff, shared with all allowlisted providers: `codeburn price-override` and `model-alias` do not affect kiro dollar amounts (token *counts* are unaffected). Historical caches from before this change re-parse via the `CACHE_VERSION` bump to 5.
|
||
- **`projectPath` for git attribution.** The parser now records the session's working directory as `projectPath` (CLI `meta.cwd`, v2 `workspacePaths[0]`, workspace sessions' `workspaceDirectory`), which sync attribution needs to resolve the git repo. The `project-path-v1` parse-version bump re-parses cached kiro history once; sessions in linked git worktrees now group under the main repo.
|
||
|
||
## When fixing a bug here
|
||
|
||
1. If the bug is "wrong workspace", check the base64 fallback path. Some users name their workspaces with characters that are not valid base64.
|
||
2. If the bug is "missing model in pricing", add the model to the normalization map and verify against `tests/providers/kiro.test.ts`.
|
||
3. If the bug is "tools missing", check both text-envelope extraction and structured tool-call extraction. Kiro changes its envelope occasionally.
|