AWRBLOCK
Persistent work state and minimal context for long-running AI agents.
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
Let your AI team keep complex projects moving—from goal to verified delivery.
The open-source project delivery platform for people and AI.
English · 简体中文
[](LICENSE) [](https://github.com/originoneai/awr/releases) [](https://www.npmjs.com/package/@originoneai/agent-work-runtime) [](https://pypi.org/project/agent-work-runtime/)
[](https://awr.originoneai.com/) [](#quickstart)
Documentation · Releases · Report an issue · Contribute
AWR connects goals, workstreams, dependencies, context and acceptance into one project delivery workflow. It is built for one person coordinating several agents, and teams delivering a shared project—starting with complex software projects.
Changing sessions, models or collaborators should preserve the project's goals, constraints and verified progress. Parallel work should have clear ownership and handoffs. At every stage, you should be able to tell what is ready, what is waiting, what has been verified, and where the effort went.
People set direction and review outcomes; agents do the work. AWR connects their work to the same project through CLI/MCP, with Team collaboration developing on that foundation. The availability
92dc5791bf53OBSERVED · 2026-10-11Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add awr-inspector --env TIKTOKEN_CACHE_DIR=${TIKTOKEN_CACHE_DIR} -- npx -y [email protected]{
"mcpServers": {
"awr-inspector": {
"command": "npx",
"args": [
"-y",
"[email protected]"
],
"env": {
"TIKTOKEN_CACHE_DIR": "${TIKTOKEN_CACHE_DIR}"
}
}
}
}Exposed tools (8)
5 read · 3 write · 0 destructive.
| Tool | Risk | Description |
|---|---|---|
awr_change | write | Preview, apply, inspect, and recover reviewed source changes. |
awr_compaction | read | Observe native compaction telemetry, read assessments, or defer advice. |
awr_context | read | Compile L1 context, prepare work or completion, and assess management. |
awr_continuity | read | Persist user waits/replies, read durable operation outcomes, reindex sources. |
awr_evidence | read | Append bounded generic events and record evidence bindings. |
awr_query | read | Read current action status, claimable work, one work item, or bounded search. |
awr_session | write | Start, inspect, list, checkpoint, end, resume, or claim bound work sessions. |
awr_work | write | Apply reviewed work transitions, manage progress, and read dependency graphs. |
Trust audit
BLOCKgrade D · trust 62/100 Do not install this without reading the findings. The audit found something that could harm you or your machine.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | FAIL |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | PASS |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (2 observation(s))
- Network
- none-observed
- Shell
- declared (1 observation(s))
- Dependencies
- pinned
- Secrets in source
- found
Findings (25)
"-----BEGIN PRIVATE KEY-----\nsynthetic material",
assert!(ensure_public_text("-----BEGIN PRIVATE KEY-----").is_err());return yaml.load(path.read_text(), Loader=UniqueLoader)
let url = "postgres://u:[email protected]:1/db?sslmode=require&connect_timeout=1";
AWR_TEAM_TEST_DATABASE_URL: postgres://postgres:[email protected]:5432/postgres
"postgres://postgres:[email protected]/awr",
export AWR_TEAM_TEST_DATABASE_URL='postgres://postgres:[email protected]:55432/postgres'
"postgres://localhost:[email protected]:55432/db",
let token = "awr1.mcp-member.2222222222222222222222222222222222222222222222222222222222222222";
let token = "awr1.mcp-member.2222222222222222222222222222222222222222222222222222222222222222";
secret = "s3cret-password-should-not-leak"
secret = "another-secret-value"
"# Protocol\n-----BEGIN PRIVATE KEY-----\nsynthetic material\n",
-----BEGIN PRIVATE KEY-----
return yaml.load(path.read_text(), Loader=UniqueLoader)
document = yaml.load(body.decode('utf-8'), Loader=UniqueLoader)PathBuf::from(env!("CARGO_MANIFEST_DIR")).join("../../tests/fixtures/assessment/replay");const SCOPED_WORK: &str = include_str!("../../../tests/fixtures/workstreams/context.yaml");const SCOPED_MANIFEST: &str = include_str!("../../../tests/fixtures/workstreams/context.toml");s["database"]["path"] = json!("../../.awr/state.db");include_str!("../../../examples/basic/project.toml"),"http://127.0.0.1:4321/callback",
"http://127.0.0.2:1234/callback",
let origin = format!("http://127.0.0.1:7381");endpoint: "http://127.0.0.1/team/v1".into(),
Gates applied: critical_finding, no_behavioural_pass.
92dc5791bf53full audit observations/trust-audit/mcp-server/originoneai__awr.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-11 | 92dc5791bf53 | BLOCK | D | 62 | first audit |
Questions
What is the @originoneai/agent-work-runtime MCP server?
Persistent work state and minimal context for long-running AI agents.
What tools does @originoneai/agent-work-runtime expose?
8 in total: 5 read-only, 3 that write, and 0 that can delete or overwrite. Every one is listed on this page with its risk.
Is @originoneai/agent-work-runtime safe to connect to an agent?
No — not without reading the findings first. The audit graded it D (62/100) and found 4 critical or high issues in the source. Each one is listed on this page with the file and line it is on.
What credentials does @originoneai/agent-work-runtime need?
It reads TIKTOKEN_CACHE_DIR from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does @originoneai/agent-work-runtime run?
It speaks streamable-http, so it runs as a service you connect to over the network. It is published on npm as awr-inspector at 1.1.0.
How current is this page?
The grade is for one exact copy of the source (92dc5791bf53), read on 2026-10-11. The repository is watched and re-audited when it changes.