@vectorize-io/hindsight-coding-agents 0.8.0 → 0.8.2
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- package/README.md +73 -8
- package/dist/antigravity-hook.js +444 -80
- package/dist/antigravity-statusline.js +9 -4
- package/dist/antigravity-stop-hook.js +444 -95
- package/dist/claude-hook.js +444 -80
- package/dist/claude-sessionstart-hook.js +476 -93
- package/dist/claude-stop-hook.js +444 -95
- package/dist/cline.js +322 -122
- package/dist/codebuddy-hook.js +4273 -0
- package/dist/codebuddy-sessionstart-hook.js +5033 -0
- package/dist/codebuddy-stop-hook.js +4567 -0
- package/dist/codex-hook.js +444 -80
- package/dist/codex-sessionstart-hook.js +476 -93
- package/dist/codex-stop-hook.js +444 -95
- package/dist/copilot-hook.js +444 -80
- package/dist/copilot-sessionstart-hook.js +476 -93
- package/dist/copilot-stop-hook.js +444 -95
- package/dist/cursor-hook.js +444 -80
- package/dist/cursor-sessionstart-hook.js +476 -93
- package/dist/cursor-stop-hook.js +444 -95
- package/dist/daemon-start.js +16 -8
- package/dist/dcode-hook.js +444 -80
- package/dist/dcode-sessionstart-hook.js +476 -93
- package/dist/dcode-stop-hook.js +444 -95
- package/dist/deepen.js +252 -79
- package/dist/devin-hook.js +444 -80
- package/dist/devin-sessionstart-hook.js +476 -93
- package/dist/devin-stop-hook.js +444 -95
- package/dist/droid-hook.js +444 -80
- package/dist/droid-sessionstart-hook.js +476 -93
- package/dist/droid-stop-hook.js +444 -95
- package/dist/dsh.js +329 -122
- package/dist/grok-hook.js +444 -80
- package/dist/grok-sessionstart-hook.js +476 -93
- package/dist/grok-stop-hook.js +444 -95
- package/dist/hindsight-seed.js +9 -4
- package/dist/index.js +331 -131
- package/dist/installer.js +1046 -647
- package/dist/kilo.js +331 -131
- package/dist/kimi-hook.js +444 -80
- package/dist/kimi-sessionstart-hook.js +476 -93
- package/dist/kimi-stop-hook.js +444 -95
- package/dist/mcp-server.js +272 -89
- package/dist/opencode2.js +328 -128
- package/dist/pi.js +384 -134
- package/dist/prime-agent.js +384 -134
- package/dist/qwen-hook.js +444 -80
- package/dist/qwen-sessionstart-hook.js +476 -93
- package/dist/qwen-stop-hook.js +444 -95
- package/dist/status.js +124 -43
- package/dist/traecode-hook.js +444 -80
- package/dist/traecode-sessionstart-hook.js +476 -93
- package/dist/traecode-stop-hook.js +444 -95
- package/dist/workbuddy-hook.js +4273 -0
- package/dist/workbuddy-sessionstart-hook.js +5033 -0
- package/dist/workbuddy-stop-hook.js +4567 -0
- package/dist/zcode-hook.js +444 -80
- package/dist/zcode-sessionstart-hook.js +476 -93
- package/dist/zcode-stop-hook.js +444 -95
- package/package.json +9 -3
- package/plugin.json +1 -1
- package/skill/SKILL.md +10 -6
package/README.md
CHANGED
|
@@ -2,7 +2,7 @@
|
|
|
2
2
|
|
|
3
3
|
Long-term project memory for **coding agents**, backed by [Hindsight](https://vectorize.io/hindsight).
|
|
4
4
|
One package, several agents: a shared reflect-and-inject core with a thin entry point per agent
|
|
5
|
-
(**Claude Code**, **Codex CLI**, **DeepAgents Dcode**, **opencode**, **opencode 2**, **Kilo CLI**, **Cursor CLI**, **GitHub Copilot CLI**, **Grok Build**, **Qwen Code**, **Kimi Code**, **Factory Droid**, **ZCode**, **TraeCode**, **Antigravity CLI**, **Devin CLI**, **Cline CLI**, **pi**, **Prime Agent**, **DeepSeek Harness**). Ingestion is fully
|
|
5
|
+
(**Claude Code**, **Codex CLI**, **DeepAgents Dcode**, **opencode**, **opencode 2**, **Kilo CLI**, **Cursor CLI**, **GitHub Copilot CLI**, **Grok Build**, **Qwen Code**, **Kimi Code**, **Factory Droid**, **ZCode**, **TraeCode**, **Antigravity CLI**, **Devin CLI**, **Cline CLI**, **pi**, **Prime Agent**, **DeepSeek Harness**, **WorkBuddy**, **CodeBuddy**). Ingestion is fully
|
|
6
6
|
automatic — there is no setup command: a repo's git history and conversations flow into its memory
|
|
7
7
|
bank in the background as you work.
|
|
8
8
|
|
|
@@ -16,6 +16,33 @@ in front of the agent at the moment it starts working, and keeps a curated set o
|
|
|
16
16
|
|
|
17
17
|
[View Changelog →](https://hindsight.vectorize.io/changelog/integrations/coding-agents)
|
|
18
18
|
|
|
19
|
+
## Does it work?
|
|
20
|
+
|
|
21
|
+
[AMB's sdebench](https://agentmemorybenchmark.ai/dataset/sdebench) is the measurement: a coding
|
|
22
|
+
agent over 61 real repository tasks, three runs each, with and without memory.
|
|
23
|
+
|
|
24
|
+
Accuracy is not what moves — every run solves 60–61 of the 61 either way. These are tasks a good
|
|
25
|
+
agent can already do. What memory changes is how much it costs you to get there: how often you have
|
|
26
|
+
to stop and correct the agent, and what the tokens cost.
|
|
27
|
+
|
|
28
|
+
<!-- figure: coding-agents-benchmark -->
|
|
29
|
+
|
|
30
|
+
| Agent | Corrections / task | Cost / task |
|
|
31
|
+
| ----------- | ------------------ | ------------------- |
|
|
32
|
+
| Claude Code | 0.85 → **0.36** | $0.445 → **$0.338** |
|
|
33
|
+
| Codex CLI | 1.34 → **0.47** | $0.577 → **$0.276** |
|
|
34
|
+
| opencode | 1.20 → **0.80** | $0.634 → **$0.553** |
|
|
35
|
+
|
|
36
|
+
No memory → with Hindsight, mean of 3 runs. Corrections reproduce the labels AMB prints on its own
|
|
37
|
+
chart; cost is derived from the same per-run table.
|
|
38
|
+
|
|
39
|
+
Both numbers move together, and that is the mechanism rather than a coincidence. A correction is a
|
|
40
|
+
round trip: the agent guesses a project-specific decision wrong, you stop it, you explain, it redoes
|
|
41
|
+
the work. Every one of those costs a turn of tokens and a turn of your attention. Handing the agent
|
|
42
|
+
the decision _before_ it starts — the rounding rule, the retry allowlist, the tie-break policy that
|
|
43
|
+
was settled in a commit message or a conversation months ago — removes the guess, and the round trip
|
|
44
|
+
with it.
|
|
45
|
+
|
|
19
46
|
## Install
|
|
20
47
|
|
|
21
48
|
<!-- skill:begin title="Install / update" -->
|
|
@@ -145,7 +172,9 @@ npx @vectorize-io/hindsight-coding-agents install grok-build
|
|
|
145
172
|
|
|
146
173
|
Native hooks in `~/.grok/hooks/hindsight.json`, MCP in `~/.grok/config.toml`, plus the companion
|
|
147
174
|
skill. Grok also runs Claude Code's hooks from `~/.claude/settings.json`; those stay silent inside
|
|
148
|
-
Grok, so a machine wired for both hosts records each Grok session once.
|
|
175
|
+
Grok, so a machine wired for both hosts records each Grok session once. Grok passes hook context to
|
|
176
|
+
the model only from tool hooks, so the session and prompt context arrive with the result of the
|
|
177
|
+
agent's first tool call after each prompt (a `PostToolUse` hook).
|
|
149
178
|
|
|
150
179
|
#### <img src="https://hindsight.vectorize.io/img/harness/qwen-code.svg" alt="" width="20" height="20" /> Qwen Code
|
|
151
180
|
|
|
@@ -334,6 +363,38 @@ too: the package ships the profile patch layer, so nothing else needs editing. E
|
|
|
334
363
|
companion skill — a plugin wired by the host's own plugin manager installs it itself on the first
|
|
335
364
|
session, since that route never runs our installer.
|
|
336
365
|
|
|
366
|
+
#### <img src="https://hindsight.vectorize.io/img/harness/workbuddy.png" alt="" width="20" height="20" /> WorkBuddy
|
|
367
|
+
|
|
368
|
+
```bash
|
|
369
|
+
npx @vectorize-io/hindsight-coding-agents install workbuddy
|
|
370
|
+
```
|
|
371
|
+
|
|
372
|
+
3 hooks in `~/.workbuddy/settings.json`, the stdio MCP server in `~/.workbuddy/mcp.json`, and the
|
|
373
|
+
companion skill in `~/.workbuddy/skills`. WorkBuddy (Tencent's AI workbench) is built on the shared
|
|
374
|
+
`@genie/agent-cli` engine, so its hooks follow Claude Code's protocol — the main difference is the
|
|
375
|
+
transcript schema (`type:"message"` records carrying top-level `role`/`content`), which the
|
|
376
|
+
package's own reader normalizes.
|
|
377
|
+
|
|
378
|
+
#### <img src="https://hindsight.vectorize.io/img/harness/codebuddy.png" alt="" width="20" height="20" /> CodeBuddy
|
|
379
|
+
|
|
380
|
+
```bash
|
|
381
|
+
npx @vectorize-io/hindsight-coding-agents install codebuddy
|
|
382
|
+
```
|
|
383
|
+
|
|
384
|
+
3 hooks in `~/.codebuddy/settings.json`, the stdio MCP server in `~/.codebuddy/.mcp.json` — the
|
|
385
|
+
recommended file: the installer follows CodeBuddy's documented fallback chain, merging into an
|
|
386
|
+
existing deprecated `~/.codebuddy/mcp.json` or legacy `~/.codebuddy.json` rather than shadowing it —
|
|
387
|
+
and the companion skill in `~/.codebuddy/skills`. CodeBuddy Code runs the same `@genie/agent-cli`
|
|
388
|
+
engine as WorkBuddy — WorkBuddy merely ships it with a different home folder — so it reuses
|
|
389
|
+
WorkBuddy's hook runtime, and differs only in the root it writes to.
|
|
390
|
+
|
|
391
|
+
**Both CodeBuddy hosts are covered, and they do not store sessions alike.** CodeBuddy Code (the CLI)
|
|
392
|
+
writes the WorkBuddy JSONL that the shared reader parses. The IDE hands the Stop hook its own
|
|
393
|
+
per-conversation directory instead — `CodeBuddyIDE/<uid>/history/<md5(workspace)>/<conversationId>/`,
|
|
394
|
+
whose `index.json` _is_ the `transcript_path` and whose prose sits one level down, in
|
|
395
|
+
`messages/<messageId>.json` — so the reader dispatches on the shape it was handed
|
|
396
|
+
(core/transcript-codebuddy-ide.ts).
|
|
397
|
+
|
|
337
398
|
Uninstall the same way: `npx @vectorize-io/hindsight-coding-agents uninstall claude-code` (or `uninstall all`).
|
|
338
399
|
|
|
339
400
|
**Devin CLI needs Node 22.5 or newer.** Its hooks pass only a session id — the conversation itself
|
|
@@ -603,12 +664,12 @@ hook by Codex...), so one shared config serves several agents side by side:
|
|
|
603
664
|
| `reflectToolTimeoutMs` | `330000` | timeout for the agent-invoked `hindsight_reflect` tool — a call the agent waits on, whose high-budget synthesis on a populated bank runs for minutes. Defaults above the server's own reflect wall timeout (`HINDSIGHT_API_REFLECT_WALL_TIMEOUT`, 300s) so the server decides when to give up. Unset, it inherits an explicitly raised `reflectTimeoutMs`, but a short one never lowers it |
|
|
604
665
|
| `injectTimeoutMs` | `7000` | retrieval timeout in milliseconds for `autoInject: "pages"` and `autoInject: "recall"`; also the **shared** budget for page search then recall after a reflect timeout/5xx (recall gets only the time left after page search). Set e.g. `2000` to limit turn latency or `15000` for a slower self-hosted bank. Environment fallback: `HINDSIGHT_INJECT_TIMEOUT_MS`. On hook harnesses, injection must fit the host's prompt-hook timeout (30s by default); for reflect with fallback, allow for **`reflectTimeoutMs` + `injectTimeoutMs`**, and raise the host hook's `timeout` when needed |
|
|
605
666
|
| `reflectBudget` | `"high"` | reflect budget for the `hindsight_reflect` tool: `"low"`, `"mid"` or `"high"`. Drop it on a large bank where high-budget synthesis exceeds the server's wall timeout. The automatic session-start reflect always uses `"low"` to fit its hook window and is unaffected |
|
|
606
|
-
| `autoInject` | `"reflect"` | what to inject once, on the session's **first prompt**: `"reflect"` = a low-budget reflect synthesis (on timeout/5xx it falls back to page search, then recall); `"pages"` = the knowledge pages matching the prompt by search; `"recall"` = the bank's memories recalled for the prompt (what it asks for is `recallOptions`, observations by default); `"none"` = nothing — the agent searches knowledge pages first and reflects only when they are too shallow. `"pages"` and `"recall"` are retrieval only (no LLM), their time budget is `injectTimeoutMs
|
|
667
|
+
| `autoInject` | `"reflect"` | what to inject once, on the session's **first prompt**: `"reflect"` = a low-budget reflect synthesis (on timeout/5xx it falls back to page search, then recall); `"pages"` = the knowledge pages matching the prompt by search; `"recall"` = the bank's memories recalled for the prompt (what it asks for is `recallOptions`, observations by default); `"none"` = nothing — the agent searches knowledge pages first and reflects only when they are too shallow. `"pages"` and `"recall"` are retrieval only (no LLM), their time budget is `injectTimeoutMs`, and when one finds nothing for the prompt it tries again on the next one, up to 5 prompts |
|
|
607
668
|
| `autoReflect` | `true` | **deprecated** — use `autoInject`. Still honoured (`false` = `autoInject: "none"`), ignored when `autoInject` is set, and logs a deprecation warning |
|
|
608
669
|
| `pageSearchLimit` | `3` | knowledge pages ONE search returns. Applies to every search of the bank — the `autoInject: "pages"` injection, the reflect fallback, and the agent's own `hindsight_search_knowledge_pages` tool — because the limit lives on the client they all share |
|
|
609
670
|
| `recallOptions` | see description | overrides merged key-by-key into the body of every recall — the `autoInject: "recall"` source and the reflect fallback. Keys are the API's own recall parameters, passed straight through, so anything recall accepts is settable without a new option here; `query` is the one field it cannot replace. The default body is `{"types": ["observation"], "budget": "low", "max_tokens": 2000, "include": {"entities": null}}`, and what you set is merged over it one key at a time — `{"max_tokens": 4000}` changes the budget and leaves the rest alone. Observations are the consolidated layer, so they answer best per token, but a bank with **consolidation disabled never grows any** and the default recall comes back empty on it: set `{"types": ["world", "experience"]}` there, or `{"types": null}` for every type. File-only, like `retainMetadata` — an object does not flatten into an env var |
|
|
610
671
|
| `toolGuideExtra` | — | your own guidance on how the agent should treat memory, added **after** the built-in tool guide (at session start and on every refresh) and after the crediting note returned with `hindsight_search_knowledge_pages` results. It adds to the built-in text, never replaces it. Set it per harness or per bank like any other key, e.g. `"Memory is a past record, not current state: verify any claim against the code before acting on it."` |
|
|
611
|
-
| `pageRefreshEveryTurns` | `
|
|
672
|
+
| `pageRefreshEveryTurns` | `1` | refetch the knowledge pages and re-inject the page roster + tool guide every N user turns (`0` also means 1) |
|
|
612
673
|
| `pageTriggerType` | `"cron"` | when NEW knowledge pages refresh, i.e. what keeping them current costs — `"cron"` (default) on `pageTriggerCron` only and only when actually stale, `"auto-refresh"` after every consolidation that produced new material, `"manual"` never on their own. Auto-refresh is the most current and by far the most expensive: one synthesis per page per consolidation. Maps to the page's `trigger.refresh_cron`, or `trigger.refresh_after_consolidation` in the Hindsight API (`true` for auto-refresh, `false` for manual) |
|
|
613
674
|
| `pageTriggerCron` | `"H * * * *"` | schedule for `pageTriggerType: "cron"` — UTC, standard 5-field cron, e.g. `"0 3 * * *"`. The default is hourly, each page on its own hashed minute. Sets the page's `trigger.refresh_cron`, which the API treats as mutually exclusive with `refresh_after_consolidation`; a scheduled refresh is skipped when nothing changed. Write a field as `H` to give each page its own value there — see **Spreading refreshes with `H`** below |
|
|
614
675
|
| `pages` | every page | per-page configuration for the seeded knowledge pages, keyed by page name (case-insensitive): `false` skips a page entirely, `{"source_query": "..."}` seeds it with your question instead of the built-in one. Omitted, all five pages are seeded with their built-in queries. This is the supported way to own a page's wording — the plugin re-syncs a page whose live query differs from the one it is configured to have, so a query edited through the API or the control plane is replaced on the next session. A skipped page is **not deleted**: one already seeded keeps its content and stops being re-synced. The scoping clause is appended to your query too, so a reworded page cannot start reporting a dependency's decisions as this project's. File-only, like `recallOptions`; in a `banks.<id>` section it replaces the global map rather than merging into it |
|
|
@@ -1121,8 +1182,8 @@ event is written on those turns — nothing reflected.
|
|
|
1121
1182
|
`~/.hindsight/coding-agents-logs/usage.jsonl` (override `HINDSIGHT_USAGE_FILE`) — the `hindsight_*`
|
|
1122
1183
|
tools the agent called during that turn, and whether its reply credited Hindsight memory ("From
|
|
1123
1184
|
Hindsight memory"). The credit rate counts only turns that called a retrieval tool (search, list,
|
|
1124
|
-
read, reflect); saving a document is not expected to be credited. Recorded
|
|
1125
|
-
`retainSessions
|
|
1185
|
+
read, reflect); saving a document is not expected to be credited. Recorded at the end of each turn,
|
|
1186
|
+
whether or not `retainSessions` writes the session back. It never leaves your machine. Summarize it per agent with:
|
|
1126
1187
|
|
|
1127
1188
|
```bash
|
|
1128
1189
|
npx @vectorize-io/hindsight-coding-agents stats
|
|
@@ -1151,8 +1212,12 @@ delete:
|
|
|
1151
1212
|
"✅ completed" once its findings land. It is retained under the `survey` strategy, whose marker
|
|
1152
1213
|
rule extracts **nothing** from a status marker, and it drives the re-survey cadence
|
|
1153
1214
|
(`surveyRefreshCommits`) and `surveyBaseline` in sync status.
|
|
1154
|
-
- `gitlog:<repo>` — the aggregated commit-message seed document,
|
|
1155
|
-
|
|
1215
|
+
- `gitlog:<repo>` — the aggregated commit-message seed document, oldest commit first. Once seeded it
|
|
1216
|
+
is kept current by **appending** only the commits since the one its `gitlog-head:<sha>` tag
|
|
1217
|
+
names, so a new commit costs the extraction of its own message, not of the whole log. It is
|
|
1218
|
+
written in full again (bounded by `seedLimit`) only when an append cannot describe HEAD: the
|
|
1219
|
+
recorded commit is no longer an ancestor of HEAD (a rebase, a force-push), more than `seedLimit`
|
|
1220
|
+
commits are new, or the server cannot append safely.
|
|
1156
1221
|
|
|
1157
1222
|
### When memory seems to be missing
|
|
1158
1223
|
|