* feat(retrieval): assemble auto-recall context server-side via /search mode="context"
Auto-recall assembly lived in every harness plugin: each one searched per
memory type, read hits back one by one, and stitched a context block with its
own budget and degradation rules. The implementations drifted, and the shared
weaknesses showed up in production injections — roughly half of the entries
degraded to a bare URI plus a score, character budgets distorted up to 6x on
CJK text, and adjacent turns re-injected the same memories.
This moves assembly into the server as one round trip. /find stays an unchanged
stateless primitive. /search gains mode="context" (mode="list" is the default
and byte-identical to before), and /recall becomes a thin preset over the same
kernel with its v1 field names folded onto the new contract.
New assembly kernel under openviking/retrieve/context_assembler/:
- Token budgeting with a CJK-aware estimate replaces the character budget.
- detail="auto" fills breadth-first then deepens: every candidate gets a
readable floor, then overview, then full for high-scoring entries. An
oversized tier falls back to the previous one instead of being truncated,
bounded by max_tokens / candidates * 2 per entry.
- Overview extraction dispatches by source: memory files use their leading
Summary section, code files reuse code_outline signatures, long documents use
a heading tree plus first paragraph.
- Directory hits start at overview and read their .overview.md sidecar, since
directories carry no stored abstract; their full tier stays capped at
overview. v1 injected the sidecar as if it were a whole file.
- Quotas generalize beyond memory types to resources and skills, with purpose
presets supplying ratios when quotas are absent.
- dedup_turns keeps a per-session ledger at {session_uri}/.recall_log.json so
every harness inherits cross-turn dedup; exclude_uris remains as the
stateless fallback.
- Rendering flattens to one <memory uri=... type=... score=... detail=...>
element per entry. Every tier carries its URI, so the model can always drill
down through the MCP read tool.
- Query expansion and digest rewriting are opt-in and fail closed: both have
timeout fuses, and a failed rewrite still returns the unrewritten block.
Retrieval failures are counted into stats rather than silently yielding an
empty block.
Plugins now send one context request, falling back to /recall and then to raw
find on older deployments, and cache that outcome so only the first turn pays
for the probe. The tri-state recallRewrite knob chooses between local host-CLI
compression and the server digest, and client-side settings move to a plugin
section in ovcli.conf.
* refactor(retrieval): give context tiers a per-category default
The tier ladder assumed `abstract` is a cheap summary. For memory files it
is not: the memory writer stores the whole stripped body in that scalar
because it doubles as the embedding text, so `abstract` costs the same as
`full` and the ladder runs `uri < overview < abstract = full`. Two of the
model's properties fell out of that: exempting `abstract` from the per-entry
cap let a single entry eat several times the budget, and `detail` — which
only ever set a ceiling — collapsed to two distinguishable behaviours across
its four values, since `auto` already allowed `full` for memory.
Tiers now come from a per-category constant table that treats the storage
shape as a given: `events` starts at overview (the one memory type whose
`# Summary` extraction is a real compression) and may deepen to full on
leftover budget; every other category is served at `abstract`, which for
memory already is the complete file at zero read cost and for resources and
skills is the generated 256-char summary. The table carries the note to move
`events` back to `abstract` once the writer stores a separate summary scalar.
Falling out of that: prefetch now reads only the candidates whose planned
tier needs a body rather than every candidate, `detail` becomes a real pin
(start and ceiling) and additionally accepts a per-category map, and
`full_score_threshold` is gone — leftover budget is spent in score order
instead of behind an absolute threshold the observed score band cannot
support. `auto` is still accepted on the wire as a synonym for "unset".
Assembly fixes found alongside:
- Removing the abstract cap exemption would turn an oversized abstract into
a bare URI, so it now falls back to overview first — for memory that is a
cheaper substitute, not a step up.
- Rewrite timeouts were reported as failures on Python 3.10, where
`asyncio.TimeoutError` is a separate class from the builtin.
- `stats.rewrite_usage` read `token_tracker` off `VLMConfig`, which has no
such attribute; usage was structurally always null. It now reads the model
instance's tracker and reports only when the call count moved by exactly
one, since that tracker is shared.
- A single malformed ledger record made every deduped recall in that session
fail, and the file was never rewritten, so it could not heal. Records are
now coerced on read and dropped on the next write, along with records left
ahead of the clock by an archive rotation.
- Entries served as a bare URI no longer enter the dedup cooldown: they lost
to budget pressure, not to the reader having already seen them.
- The render envelope only neutralised a literal `</memory>`, so a body could
forge a sibling entry with its own uri, type and score.
- Flat-mode gathering re-derived the category from the URI, reading
`viking://resources/backup/memories/events/log.md` as an event.
- Cooled and excluded URIs are compensated with extra rows, so a fully cooled
bucket falls through to the next-best hits instead of coming back empty.
- `/recall` quotas overlay the v1 bucket defaults again; `{"events": 5}` had
started dropping the other three buckets.
- The MCP `recall` signature sent its own defaults as if the caller had, which
resolved a different profile than `POST /recall`; an unknown `detail` value
raised `KeyError` through the whole call instead of degrading.
* feat(codex): inject profile context on session start
Reuse the shared profile builder for startup, clear, and resume hooks while preserving archive injection and orphan-session status output.
Co-authored-by: TRAE CLI <noreply@bytedance.com>
* fix(retrieval): raise rewrite timeout default to 30s
* docs(agents): document low-latency recall settings
* fix(codex): prefer luna as recall compressor fallback
* refactor(plugins): unify recall compression setting
* feat(plugins): enable recall compression by default
* docs(agents): use absolute links in image docs
* fix(retrieval): address context assembly review feedback
Co-authored-by: TRAE CLI <noreply@bytedance.com>
* test: trim redundant context assembly coverage
* fix(retrieval): address second-round context assembly review
- Drop the backticked `/search` from the deprecated-recall row in both API
overviews. The reference checker scans the whole row after the method cell
for backticked paths, so it read the description as a route named
`POST /search` and Build Docs failed on an unknown, undocumented route.
- Accept ovcli.conf's full field set in both Python readers. The file's schema
belongs to the Rust CLI, which writes `root_api_key`, `output`,
`echo_command`, `show_progress` and `verbose` and ignores unknown keys; the
two Python readers had drifted into stricter subsets, so the shipped example
already failed to load in both. Adding the new `plugin` section to a working
ovcli.conf would have broken `ov doctor` and every SDK client the same way.
- Return 400 from `mode="context"` for a request `mode="list"` also rejects.
Retrieval validates query and image_url before searching, and the gather
fuse swallowed that rejection along with genuine scope failures, so a body
of `{"mode":"context"}` came back 200 with an empty block instead of the
documented parameter error. Runtime failures still degrade into
`stats.retrieval_errors`.
- Let a context request that asks for a server-side digest outlast the
server's rewrite fuse. The plugin's ordinary 15s request timeout is shorter
than the 30s fuse, so a rewrite that finished inside its own budget was
aborted client-side, discarding the whole response — including the
uncompressed block the server returns when a rewrite fails — and falling
back to `/recall`. The deadline is only extended when the body actually
requests a rewrite, and `OPENVIKING_RECALL_CONTEXT_TIMEOUT_MS` /
`plugin.recallContextTimeoutMs` pins it.
* chore(plugins): sync shared modules into the zcode snapshot
* fix(retrieval): align context quotas and plugin defaults
Restore cross-domain coding recall, reuse authoritative actor resource
scopes, and make bucket quotas the sole width control in purpose mode.
Keep plugin defaults server-owned while preserving explicit legacy limit
settings through quota conversion.
Co-authored-by: TRAE CLI <noreply@bytedance.com>
* fix(retrieval): preserve recall compatibility
Restore the deprecated recall threshold default, distinguish successful empty rewrites from compressor failures, and document legacy quota floors across coding-agent plugins.
Co-authored-by: TRAE CLI <noreply@bytedance.com>
---------
Co-authored-by: TRAE CLI <noreply@bytedance.com>
Co-authored-by: qin-ctx <qinhaojie.exe@bytedance.com>
7.5 KiB
OpenViking OpenCode Plugin
A unified OpenCode plugin for OpenViking repository retrieval and long-term memory.
This is the only OpenCode plugin example maintained in this repository. It supersedes the former split examples for indexed repository prompt injection and long-term memory.
The plugin uses OpenCode hooks for lifecycle behavior and registers OpenViking's standard stdio MCP proxy for model tools. It does not install or require an OpenCode skill, and agents do not need to run ov shell commands.
What It Does
- Injects indexed
viking://resources/repositories into the system prompt. - Exposes the same OpenViking MCP tools used by the Claude Code and Codex memory plugins.
- Maps each OpenCode session to an OpenViking session.
- Captures user and assistant text messages into OpenViking.
- Commits sessions at lifecycle boundaries for memory extraction.
- Automatically recalls relevant memories and injects them as hidden synthetic context for the current user message.
- Blocks accidental local filesystem reads of
viking://URIs and points the agent back toopenviking_read,openviking_glob, oropenviking_search.
Files
examples/opencode-plugin/
├── index.mjs
├── package.json
├── README.md
├── INSTALL-ZH.md
├── lib/
│ ├── config.mjs
│ ├── mcp-config.mjs
│ ├── runtime.mjs
│ ├── repo-context.mjs
│ ├── memory-session.mjs
│ ├── memory-recall.mjs
│ ├── session-inject.mjs
│ ├── viking-uri-guard.mjs
│ └── utils.mjs
├── servers/
│ └── mcp-proxy.mjs
├── tests/
└── wrappers/
└── openviking.js
There is intentionally no skills/openviking/SKILL.md. The tool surface comes from OpenViking's MCP endpoint.
Requirements
- OpenCode
- OpenViking HTTP server
- Node.js 18+
- An OpenViking API key if your server requires authentication
Start OpenViking first:
openviking-server --config ~/.openviking/ov.conf
Installation
Published Package
Normal users should enable it through OpenCode's package plugin mechanism:
The published npm package is @openviking/opencode-plugin; verify availability with:
npm view @openviking/opencode-plugin version
{
"plugin": ["@openviking/opencode-plugin"]
}
Source Install
For development or PR testing, copy the package into OpenCode's plugin directory with a top-level wrapper:
mkdir -p ~/.config/opencode/plugins/openviking
cp examples/opencode-plugin/wrappers/openviking.js ~/.config/opencode/plugins/openviking.js
cp examples/opencode-plugin/index.mjs examples/opencode-plugin/package.json ~/.config/opencode/plugins/openviking/
cp -r examples/opencode-plugin/lib ~/.config/opencode/plugins/openviking/
cp -r examples/opencode-plugin/servers ~/.config/opencode/plugins/openviking/
This creates a stable OpenCode plugin layout:
~/.config/opencode/plugins/
├── openviking.js
└── openviking/
├── index.mjs
├── package.json
├── lib/
└── servers/
The top-level openviking.js is only a wrapper:
export { OpenVikingPlugin, default } from "./openviking/index.mjs"
This wrapper is only for source installs with the directory layout shown above. npm package installs load index.mjs directly through package.json.
Use the .js wrapper for source installs; OpenCode's local plugin scanner discovers JavaScript/TypeScript plugin files.
Configuration
Create ~/.config/opencode/openviking-config.json:
{
"enabled": true,
"timeoutMs": 30000,
"repoContext": { "enabled": true, "cacheTtlMs": 60000 },
"autoRecall": {
"enabled": true,
"limit": 6,
"scoreThreshold": 0.35,
"maxContentChars": 500,
"preferAbstract": true,
"tokenBudget": 2000,
"minQueryLength": 3
},
"commitTokenThreshold": 20000,
"commitKeepRecentCount": 10,
"profileTokenBudget": 10000,
"resumeContextBudget": 32000
}
autoRecall.limit is a legacy quota-scaling input, not a final result cap.
Explicit values from 1 through 5 produce an effective total quota of 6 because
each coding category keeps one retrieval slot. Use Context quotas directly
when exact category ceilings are required.
API keys are resolved from environment variables or ~/.openviking/ovcli.conf and sent as Authorization: Bearer ... by both hooks and the MCP proxy. Recall goes through the server-side context face (POST /api/v1/search/search with mode="context"), falling back to the deprecated /api/v1/search/recall on older deployments. account and user are trusted-mode identity
headers sent as X-OpenViking-Account and X-OpenViking-User; leave them empty
when using API-key mode with user/admin API keys.
By default the plugin derives a peer from the project directory using Claude's
project-directory naming rule: every non-letter-or-digit character becomes -,
with no path normalization. For example, /Users/x/Dev/OpenViking becomes
-Users-x-Dev-OpenViking. Data-plane memory/resource requests send the
effective peer as X-OpenViking-Actor-Peer; captured session messages store it
as body peer_id. Configure peerId or OPENVIKING_PEER_ID to override the
workspace-derived peer, or set workspacePeer=false /
OPENVIKING_WORKSPACE_PEER=0 to turn workspace-derived peers off.
Recall defaults to the broad mode: global memory, the current workspace, and
other workspace memories can all be recalled, with other workspaces penalized
and rendered later. Set recallPeerScope="actor" or
OPENVIKING_RECALL_PEER_SCOPE=actor for the isolation mode, which only sees
global memory plus the current workspace. In deployments where one bot serves
multiple real people, such as zouk, vikingbot, or AstrBot, use the isolation mode
with an explicit actor peer so one person's memories are not recalled into
another person's session.
OPENVIKING_API_KEY, OPENVIKING_ACCOUNT, OPENVIKING_USER,
and OPENVIKING_PEER_ID take precedence over values in this file.
For advanced setups, OPENVIKING_PLUGIN_CONFIG can point to another config file path.
OpenCode's local read, glob, and grep tools cannot read viking:// URIs.
When the agent accidentally tries that, the plugin blocks the filesystem tool
call and points it to the OpenViking MCP tools.
MCP Tools
OpenCode sees the OpenViking MCP server as openviking, so tool names are namespaced with openviking_.
openviking_recall: balanced current-task recall using OpenViking's/recallendpoint.openviking_search: deep semantic retrieval across memories, resources, and skills.openviking_find: fast semantic retrieval.openviking_remember: store important facts or decisions for memory extraction.openviking_read: read one or moreviking://files.openviking_list: list aviking://directory.openviking_grep: exact text or regex search.openviking_glob: glob file matching.openviking_add_resource: add a URL, local file, sitemap, or feed.openviking_forget: delete aviking://URI after explicit user confirmation.openviking_list_watches/openviking_cancel_watch: inspect or cancel resource watches.openviking_health: check OpenViking server health.
The proxy forwards the server's real tools/list response; the plugin does not maintain a separate native tool list.
Runtime Files
The plugin writes runtime files to ~/.config/opencode/openviking/ by default:
openviking-memory.logopenviking-session-state.json
Set runtime.dataDir in config to override this directory.