agent-inspect 6.7.0 → 6.7.2

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/CHANGELOG.md CHANGED
@@ -1,5 +1,19 @@
1
1
  # Changelog
2
2
 
3
+ ## Unreleased
4
+
5
+ ## 6.7.2
6
+
7
+ ### Patch Changes
8
+
9
+ - 9c1f54c: Product presentation patch: rebuild README/docs/package guides/website for the 6.7.x launch candidate, fix public CLI command targets, add docs/link/public-truth validation, and slim the public roadmap. Docs and presentation only — no runtime product expansion.
10
+
11
+ ## 6.7.1
12
+
13
+ ### Patch Changes
14
+
15
+ - dea3d91: Release tooling and public-truth patch: Changesets fixed group for all 18 public packages, linked-versions CI check, README/ROADMAP aligned to 6.7.x launch candidate, MCP server initialize version from package metadata.
16
+
3
17
  ## 6.7.0
4
18
 
5
19
  ### Minor Changes
@@ -193,17 +207,15 @@
193
207
 
194
208
  - 70f3fb2: v3.1 adoption train: public `@agent-inspect/harness`, `agent-inspect init` and `doctor` commands, adoption starters, and onboarding docs.
195
209
 
196
- ## Unreleased
197
-
198
- README adoption polish (brand header, product-loop visual, npm link hygiene). Docs-only; no runtime feature changes.
210
+ ## Historical (pre-3.5 train notes)
199
211
 
200
- See [docs/implementation/reviews/README-ADOPTION-POLISH-REVIEW.md](docs/implementation/reviews/README-ADOPTION-POLISH-REVIEW.md).
212
+ The v3.0→v3.5 feature train is complete. Older in-progress notes below are kept for history.
201
213
 
202
- ---
214
+ ### README adoption polish (shipped in 3.5.x docs)
203
215
 
204
- ## Historical (pre-3.5 train notes)
216
+ README adoption polish (brand header, product-loop visual, npm link hygiene). Docs-only; no runtime feature changes.
205
217
 
206
- The v3.0→v3.5 feature train is complete. Older in-progress notes below are kept for history.
218
+ See [docs/implementation/reviews/README-ADOPTION-POLISH-REVIEW.md](docs/implementation/reviews/README-ADOPTION-POLISH-REVIEW.md).
207
219
 
208
220
  ### v3.1 (shipped in 3.1.0)
209
221
 
@@ -257,9 +269,9 @@ The v3.0→v3.5 feature train is complete. Older in-progress notes below are kep
257
269
 
258
270
  Adds the public optional `@agent-inspect/vitest` and `@agent-inspect/jest` reporter packages, the shared experimental `agent-inspect/reporters` helpers, and the `agent-inspect ci-summary` workflow for deterministic local reporter manifests and CI artifacts.
259
271
 
260
- ## Unreleased
272
+ ### Draft notes (superseded by 2.3.0 / 2.4.0 releases)
261
273
 
262
- ### Draft v2.4.0 Notes
274
+ #### Draft v2.4.0 Notes
263
275
 
264
276
  - Added session/workflow causality model and `agent-inspect/advanced` session index helpers (`buildSessionIndex`, scope/cohort helpers, session fixtures).
265
277
  - Added `agent-inspect sessions` and `agent-inspect session` CLI for multi-run handoff/retry navigation with timeline, critical-path, diagnostics, and JSON output.
@@ -267,7 +279,7 @@ The v3.0→v3.5 feature train is complete. Older in-progress notes below are kep
267
279
  - Added public optional `@agent-inspect/mcp` for local MCP **client** `tools/list` and `tools/call` telemetry with bounded summaries and `source.type: mcp-client` metadata.
268
280
  - No schema version change, no MCP gateway/server, no timestamp-only causality inference, and no root/core dependency on MCP SDKs.
269
281
 
270
- ### Draft v2.3.0 Notes
282
+ #### Draft v2.3.0 Notes
271
283
 
272
284
  - Hardened the official adapter paths for AI SDK, OpenAI Agents JS, and LangChain/LangGraph with no-network recipes, local-only defaults, clearer lifecycle coverage, and executable adapter conformance evidence.
273
285
  - AI SDK coverage now includes route-style telemetry factory guidance, per-request integration isolation, tool/stream/error/parallel fixtures, token metadata, and the required `recordInputs: false` / `recordOutputs: false` host settings.
package/README.md CHANGED
@@ -8,11 +8,16 @@
8
8
  <h1 align="center">agent-inspect</h1>
9
9
 
10
10
  <p align="center">
11
- <strong>Trace, check, and safely share TypeScript AI-agent runs locally.</strong>
11
+ <strong>Debug, regression-test, and safely share TypeScript AI-agent behavior—locally.</strong>
12
12
  </p>
13
13
 
14
14
  <p align="center">
15
- <sub>No account · no upload · no hosted dashboard · metadata-only by default</sub>
15
+ <sub>No account · no default upload · metadata-only by default · optional customer-owned Studio</sub>
16
+ </p>
17
+
18
+ <p align="center">
19
+ AgentInspect turns agent runs into <strong>customer-owned evidence</strong>: execution trees,
20
+ deterministic contracts, CI gates, verified-safe bundles, and optional self-hosted review.
16
21
  </p>
17
22
 
18
23
  <p align="center">
@@ -29,141 +34,191 @@
29
34
  <a href="https://www.typescriptlang.org/"><img src="https://img.shields.io/badge/TypeScript-strict-3178c6" alt="TypeScript"></a>
30
35
  </p>
31
36
 
32
- agent-inspect turns AI-agent runs into readable **local execution trees**: framework events, observed objects, tool calls, LLM steps, retries, failures, timings, sessions, and CI artifacts — without an account, collector, or hosted dashboard.
33
-
34
- **Default loop:** capture locally → inspect / report / diff → check in CI → redact before sharing.
35
-
36
37
  ```bash
37
38
  npm install agent-inspect
38
39
  ```
39
40
 
40
- ## 60-second quickstart
41
+ ## Three workflows
42
+
43
+ | Workflow | What you do |
44
+ | -------- | ----------- |
45
+ | **Debug one run** | Capture/import → tree / timeline / report → first causal failure |
46
+ | **Prevent one regression** | TraceContract / `check` → suite / cohort → CI gate |
47
+ | **Share one safe artifact** | Redact → `verify-safe` → offline bundle → optional Studio review |
48
+
49
+ **Review as a team (optional):** workspace + optional SQLite index → customer-owned [Studio Beta](https://github.com/rajudandigam/agent-inspect/tree/main/packages/studio) (no AgentInspect-hosted cloud).
50
+
51
+ <p align="center">
52
+ <img src="https://raw.githubusercontent.com/rajudandigam/agent-inspect/main/docs/assets/readme-product-loop.svg?sanitize=true" alt="Capture or import → understand → enforce → verify and bundle → review locally or in customer-owned Studio" width="900">
53
+ </p>
54
+
55
+ ## Five-minute path
56
+
57
+ Commands below match the packed quickstart. Replace `<run-id>` with a value from `list`.
41
58
 
42
59
  ```bash
60
+ npm install agent-inspect
43
61
  npx agent-inspect init --yes
44
62
  node examples/agent-inspect-demo.mjs
45
63
  npx agent-inspect list --dir .agent-inspect
64
+ ```
65
+
66
+ ```bash
67
+ # After copying a run id from list:
46
68
  npx agent-inspect view <run-id> --dir .agent-inspect
47
69
  npx agent-inspect report <run-id> --dir .agent-inspect
70
+ npx agent-inspect check <run-id> --dir .agent-inspect
71
+ npx agent-inspect bundle <run-id> --dir .agent-inspect --profile share
48
72
  npx agent-inspect verify-safe <run-id> --dir .agent-inspect
49
73
  ```
50
74
 
51
- Install one trace one failure check one share-safe artifact in under five minutes.
52
- Guide: [First trace in 5 minutes](https://agentinspect.vercel.app/docs/getting-started/) · [repo](docs/FIRST-TRACE-IN-5-MINUTES.md)
75
+ `init` scaffolds config and a demo script; the **demo** writes the trace. Guide: [First trace in 5 minutes](https://agentinspect.vercel.app/docs/getting-started/) · [repo](https://github.com/rajudandigam/agent-inspect/blob/main/docs/FIRST-TRACE-IN-5-MINUTES.md) · [Golden path](https://github.com/rajudandigam/agent-inspect/blob/main/docs/GOLDEN-PATH.md)
53
76
 
54
- <p align="center">
55
- <img src="https://raw.githubusercontent.com/rajudandigam/agent-inspect/main/docs/assets/readme-product-loop.svg?sanitize=true" alt="Capture, inspect, check, redact — local JSONL only" width="720">
56
- </p>
57
-
58
- ## Choose your path
77
+ ## Choose your capture path
59
78
 
60
79
  | Path | Use when | Start |
61
80
  | ---- | -------- | ----- |
62
- | **AI SDK** | Vercel AI SDK `generateText` / `streamText` | [`@agent-inspect/ai-sdk`](packages/ai-sdk/README.md) · [guide](docs/AI-SDK-ADOPTION.md) |
63
- | **OpenAI Agents** | OpenAI Agents JS | [`@agent-inspect/openai-agents`](packages/openai-agents/README.md) · [guide](docs/OPENAI-AGENTS-LOCAL.md) |
64
- | **LangChain** | Callbacks / LangGraph-via-LangChain | [`@agent-inspect/langchain`](packages/langchain/README.md) |
65
- | **Observe** | Object/class with `run` / `execute` / `invoke` | [Getting started](docs/GETTING-STARTED.md) |
66
- | **Manual** | Custom spans and nesting | `inspectRun` + `step` in [API](docs/API.md) |
67
- | **Logs** | Structured logs already emitted | [Log-to-tree](docs/LOG-TO-TREE-QUICKSTART.md) |
68
- | **CI / tests** | Failed-test artifacts | [`@agent-inspect/vitest`](packages/vitest/README.md) · [`@agent-inspect/jest`](packages/jest/README.md) |
69
- | **Real projects** | Fixture runner / bootstrap | [`@agent-inspect/harness`](packages/harness/README.md) · [NestJS](docs/NESTJS.md) |
81
+ | **Manual / observe** | Custom nesting or object methods | [Getting started](https://github.com/rajudandigam/agent-inspect/blob/main/docs/GETTING-STARTED.md) |
82
+ | **AI SDK** | Vercel AI SDK `generateText` / `streamText` | [`@agent-inspect/ai-sdk`](https://www.npmjs.com/package/@agent-inspect/ai-sdk) |
83
+ | **OpenAI Agents** | OpenAI Agents JS | [`@agent-inspect/openai-agents`](https://www.npmjs.com/package/@agent-inspect/openai-agents) |
84
+ | **LangChain** | Callbacks / LangGraph-via-LangChain | [`@agent-inspect/langchain`](https://www.npmjs.com/package/@agent-inspect/langchain) |
85
+ | **Structured logs** | Logs already emitted | [Log-to-tree](https://github.com/rajudandigam/agent-inspect/blob/main/docs/LOG-TO-TREE-QUICKSTART.md) |
86
+ | **Harness** | Fixture runner for real projects | [`@agent-inspect/harness`](https://www.npmjs.com/package/@agent-inspect/harness) |
87
+ | **CI reporters** | Failed-test artifacts | [`vitest`](https://www.npmjs.com/package/@agent-inspect/vitest) · [`jest`](https://www.npmjs.com/package/@agent-inspect/jest) |
88
+ | **Standards files** | OpenInference / OTLP JSON | [Standards](https://github.com/rajudandigam/agent-inspect/blob/main/docs/STANDARDS.md) |
70
89
 
71
90
  Blessed starters (no API keys): [examples/starters](https://github.com/rajudandigam/agent-inspect/tree/main/examples/starters)
72
91
 
73
- ## What it helps with
92
+ ## What you can do after capture
93
+
94
+ **Understand** — `what` / `view` / tree · timeline · report · diff · sessions / activity
95
+
96
+ **Prevent regressions** — deterministic checks · typed TraceContract (Beta) · suites · cohorts · CI gates · Vitest/Jest reporters (artifact reporters today; TraceContract matchers not shipped)
97
+
98
+ **Share safely** — redaction profiles · `scan` · `verify-safe` · offline bundles · CI artifacts
99
+
100
+ **Scale locally** — workspace · optional SQLite index (Beta) · observed outcomes · viewer / TUI / VS Code
74
101
 
75
- - **Wrong tool call** — tool steps, args metadata, and parent run in one tree
76
- - **Failed eval / test** — `check`, `eval`, Vitest/Jest reporters on failure
77
- - **Baseline vs candidate** local `diff` of two runs
78
- - **PR trace artifact** — [CI artifacts](docs/CI-ARTIFACTS.md) + `redact --profile share`
79
- - **Safe incident handoff** [safe sharing](https://agentinspect.vercel.app/docs/safe-sharing/) before Slack or GitHub
80
- - **Multi-agent / sessions** — `sessions`, `search`, handoff metadata
81
- - **MCP tool calls** — [`@agent-inspect/mcp`](packages/mcp/README.md) client tracing
82
- - **VS Code review** — in-repo extension ([dev guide](docs/VSCODE.md); Marketplace not published yet)
83
- - **Existing logs** — parse structured logs when you cannot instrument
102
+ **Review as a team** — customer-owned Studio Beta · explicit file / GitHub / HTTP ingest (disabled by default) · no maintainer cloud
103
+
104
+ **Interoperate** — read-only MCP server (Preview) · OpenInference-compatible / OTLP GenAI-aligned mapping with known-loss reporting
105
+
106
+ Support labels: [SUPPORT-LEVELS.md](https://github.com/rajudandigam/agent-inspect/blob/main/docs/SUPPORT-LEVELS.md) · Network: [NETWORK-BEHAVIOR.md](https://github.com/rajudandigam/agent-inspect/blob/main/docs/NETWORK-BEHAVIOR.md)
84
107
 
85
108
  ## Real-world scenarios
86
109
 
87
- | Scenario | Where to start |
88
- | -------- | -------------- |
89
- | Local debugging | [Use cases](docs/USE-CASES.md) · [broken-agent starter](https://github.com/rajudandigam/agent-inspect/tree/main/examples/starters/broken-agent-debugging) |
90
- | CI failure review | [CI artifacts](docs/CI-ARTIFACTS.md) · [ci-eval-redact](https://github.com/rajudandigam/agent-inspect/tree/main/examples/starters/ci-eval-redact) |
91
- | Team adoption | [Team workflows](docs/TEAM-WORKFLOWS.md) · [Design partners](docs/DESIGN-PARTNER-GUIDE.md) |
92
- | Safe sharing | [Safe sharing](https://agentinspect.vercel.app/docs/safe-sharing/) · [repo](docs/SAFE-TRACE-SHARING.md) |
93
- | Framework-native tracing | [Adapters](docs/ADAPTERS.md) · package READMEs above |
94
- | Design partner trial | [Design partner guide](docs/DESIGN-PARTNER-GUIDE.md) · [Demo script](docs/DEMO-SCRIPT.md) |
110
+ | Scenario | Start |
111
+ | -------- | ----- |
112
+ | Wrong tool call | [broken-agent starter](https://github.com/rajudandigam/agent-inspect/tree/main/examples/starters/broken-agent-debugging) |
113
+ | CI trajectory gate | [CI artifacts](https://github.com/rajudandigam/agent-inspect/blob/main/docs/CI-ARTIFACTS.md) |
114
+ | Safe incident handoff | [Safe sharing](https://github.com/rajudandigam/agent-inspect/blob/main/docs/SAFE-TRACE-SHARING.md) |
115
+ | Multi-agent / session retry | [Sessions & outcomes](https://github.com/rajudandigam/agent-inspect/blob/main/docs/SESSIONS-AND-OUTCOMES.md) |
116
+ | Customer-owned team review | [Self-hosting](https://github.com/rajudandigam/agent-inspect/blob/main/docs/SELF-HOSTING.md) · Studio |
117
+ | Design partner trial | [Pilot kit](https://github.com/rajudandigam/agent-inspect/blob/main/docs/PRE-V7-PILOT-KIT.md) |
118
+
119
+ ## Safety and network behavior
120
+
121
+ - Traces are **local JSONL** under `.agent-inspect/` (or `AGENT_INSPECT_TRACE_DIR`)
122
+ - **Metadata-only by default** — no raw prompts/outputs unless you opt in
123
+ - **No hidden upload** — core does not send traces to AgentInspect
124
+ - **Customer-owned Studio ingestion** is disabled by default and explicit when enabled
125
+ - **MCP server** exposes configured local evidence to the connected client (Preview)
126
+ - **Standards export** only when you run/configure it
127
+ - Redaction is **best-effort**, not certification — review before posting
128
+ - **Not** a chain-of-thought recorder
129
+
130
+ Details: [Safe sharing](https://github.com/rajudandigam/agent-inspect/blob/main/docs/SAFE-TRACE-SHARING.md) · [Network behavior](https://github.com/rajudandigam/agent-inspect/blob/main/docs/NETWORK-BEHAVIOR.md) · [Security](SECURITY.md)
131
+
132
+ ## Project status
133
+
134
+ **Current release:** **6.7.2** (eighteen linked npm packages). Technical launch candidate; external pilot evidence pending. Persisted schema **1.0**. Node.js **≥ 20**. **v7 not scheduled.**
135
+
136
+ [Roadmap](ROADMAP.md) · [Pilot kit](https://github.com/rajudandigam/agent-inspect/blob/main/docs/PRE-V7-PILOT-KIT.md) · [Changelog](CHANGELOG.md)
137
+
138
+ ## What AgentInspect is not
95
139
 
96
- ## Package map
140
+ - Hosted SaaS or maintainer-hosted dashboard
141
+ - Production APM replacement
142
+ - Eval dataset platform or LLM-as-judge by default
143
+ - Prompt registry, pricing engine, or replay engine
144
+ - Universal standards exporter or completed external adoption proof
145
+
146
+ See [Compare](https://github.com/rajudandigam/agent-inspect/blob/main/docs/COMPARE.md).
147
+
148
+ <details>
149
+ <summary><strong>Package family (18 public packages)</strong></summary>
150
+
151
+ ### Framework adapters
97
152
 
98
153
  | Package | Purpose |
99
154
  | ------- | ------- |
100
155
  | [`agent-inspect`](https://www.npmjs.com/package/agent-inspect) | Core APIs + CLI |
101
- | [`@agent-inspect/ai-sdk`](packages/ai-sdk/README.md) | AI SDK telemetry |
102
- | [`@agent-inspect/openai-agents`](packages/openai-agents/README.md) | OpenAI Agents processor |
103
- | [`@agent-inspect/langchain`](packages/langchain/README.md) | LangChain callbacks |
104
- | [`@agent-inspect/harness`](packages/harness/README.md) | Fixture runner for real projects |
105
- | [`@agent-inspect/redact`](packages/redact/README.md) | Deterministic redaction |
106
- | [`@agent-inspect/eval`](packages/eval/README.md) | Local eval heuristics |
107
- | [`@agent-inspect/vitest`](packages/vitest/README.md) | Vitest reporter |
108
- | [`@agent-inspect/jest`](packages/jest/README.md) | Jest reporter |
109
- | [`@agent-inspect/mcp`](packages/mcp/README.md) | MCP client tracing |
110
- | [`@agent-inspect/mcp-server`](packages/mcp-server/README.md) | Read-only trace MCP server |
111
- | [`@agent-inspect/guardrails`](packages/guardrails/README.md) | Deterministic guardrail rules |
112
- | [`@agent-inspect/circuit`](packages/circuit/README.md) | Loop / retry / timeout analyzers |
113
- | [`@agent-inspect/viewer`](packages/viewer/README.md) | Localhost viewer |
114
- | [`@agent-inspect/adapter-sdk`](packages/adapter-sdk/README.md) | Third-party adapters |
115
- | [`@agent-inspect/tui`](packages/tui/README.md) | Optional terminal UI |
116
- | `agent-inspect-vscode` | VS Code extension (in-repo; not on Marketplace yet) |
117
-
118
- ## Safety model
156
+ | [`@agent-inspect/ai-sdk`](https://www.npmjs.com/package/@agent-inspect/ai-sdk) | AI SDK telemetry |
157
+ | [`@agent-inspect/openai-agents`](https://www.npmjs.com/package/@agent-inspect/openai-agents) | OpenAI Agents processor |
158
+ | [`@agent-inspect/langchain`](https://www.npmjs.com/package/@agent-inspect/langchain) | LangChain callbacks |
119
159
 
120
- - Traces are **local JSONL** under `.agent-inspect/` (or `AGENT_INSPECT_TRACE_DIR`)
121
- - **Metadata-only by default** — no raw prompts/outputs unless you opt in
122
- - **No hidden upload** — AgentInspect does not send traces to the cloud
123
- - **Redaction profiles** — `local` / `share` / `strict` via CLI or [`@agent-inspect/redact`](packages/redact/README.md)
124
- - **`scan` / `verify-safe`** — check artifacts before sharing
125
- - **Not** a chain-of-thought recorder or compliance engine — review exports before posting
160
+ ### Testing / evaluation
126
161
 
127
- Details: [Safe sharing](https://agentinspect.vercel.app/docs/safe-sharing/) · [repo](docs/SAFE-TRACE-SHARING.md) · [Security](SECURITY.md)
162
+ | Package | Purpose |
163
+ | ------- | ------- |
164
+ | [`@agent-inspect/harness`](https://www.npmjs.com/package/@agent-inspect/harness) | Fixture runner |
165
+ | [`@agent-inspect/eval`](https://www.npmjs.com/package/@agent-inspect/eval) | Local eval heuristics |
166
+ | [`@agent-inspect/vitest`](https://www.npmjs.com/package/@agent-inspect/vitest) | Vitest reporter |
167
+ | [`@agent-inspect/jest`](https://www.npmjs.com/package/@agent-inspect/jest) | Jest reporter |
128
168
 
129
- ## Documentation
169
+ ### Safety
130
170
 
131
- | | Website | Repo |
132
- | - | ------- | ---- |
133
- | Getting started | [docs/getting-started](https://agentinspect.vercel.app/docs/getting-started/) | [GETTING-STARTED.md](docs/GETTING-STARTED.md) |
134
- | Safe sharing | [docs/safe-sharing](https://agentinspect.vercel.app/docs/safe-sharing/) | [SAFE-TRACE-SHARING.md](docs/SAFE-TRACE-SHARING.md) |
135
- | Compare | [docs/compare](https://agentinspect.vercel.app/docs/compare/) | [COMPARE.md](docs/COMPARE.md) |
136
- | API / CLI | — | [API.md](docs/API.md) · [CLI.md](docs/CLI.md) |
137
- | Adoption | — | [ADOPTION.md](docs/ADOPTION.md) · [USE-CASES.md](docs/USE-CASES.md) |
138
- | Technical guide | — | [TECHNICAL-GUIDE.md](docs/TECHNICAL-GUIDE.md) |
139
- | Examples | — | [starters](https://github.com/rajudandigam/agent-inspect/tree/main/examples/starters) |
140
- | Visual demos | — | [SCREENSHOTS.md](docs/SCREENSHOTS.md) |
171
+ | Package | Purpose |
172
+ | ------- | ------- |
173
+ | [`@agent-inspect/redact`](https://www.npmjs.com/package/@agent-inspect/redact) | Deterministic redaction |
174
+ | [`@agent-inspect/guardrails`](https://www.npmjs.com/package/@agent-inspect/guardrails) | Deterministic guardrail rules |
175
+ | [`@agent-inspect/circuit`](https://www.npmjs.com/package/@agent-inspect/circuit) | Loop / retry / timeout analyzers |
141
176
 
142
- Full index: [docs/README.md](docs/README.md)
177
+ ### Developer surfaces
143
178
 
144
- ## What AgentInspect is not
179
+ | Package | Purpose |
180
+ | ------- | ------- |
181
+ | [`@agent-inspect/viewer`](https://www.npmjs.com/package/@agent-inspect/viewer) | Localhost viewer |
182
+ | [`@agent-inspect/tui`](https://www.npmjs.com/package/@agent-inspect/tui) | Optional terminal UI |
183
+ | [`@agent-inspect/mcp`](https://www.npmjs.com/package/@agent-inspect/mcp) | MCP client tracing |
184
+ | [`@agent-inspect/mcp-server`](https://www.npmjs.com/package/@agent-inspect/mcp-server) | Read-only MCP server (Preview) |
185
+
186
+ ### Team / self-hosted
187
+
188
+ | Package | Purpose |
189
+ | ------- | ------- |
190
+ | [`@agent-inspect/index-sqlite`](https://www.npmjs.com/package/@agent-inspect/index-sqlite) | Optional SQLite index (Beta) |
191
+ | [`@agent-inspect/studio`](https://www.npmjs.com/package/@agent-inspect/studio) | Customer-owned Studio (Beta) |
145
192
 
146
- - Hosted SaaS or dashboard product
147
- - Production APM replacement (use LangSmith, Langfuse, OTel, etc. alongside)
148
- - Eval dataset platform or LLM-as-judge service
149
- - Prompt registry or provider pricing engine
150
- - Default telemetry uploader or replay engine
193
+ ### Extension / interop
151
194
 
152
- See [Compare](https://agentinspect.vercel.app/docs/compare/).
195
+ | Package | Purpose |
196
+ | ------- | ------- |
197
+ | [`@agent-inspect/adapter-sdk`](https://www.npmjs.com/package/@agent-inspect/adapter-sdk) | Third-party adapters (Beta) |
153
198
 
154
- ## Install details
199
+ `agent-inspect-vscode` is in-repo (Marketplace not published yet).
155
200
 
156
- Current release: **6.4.0** (sixteen linked npm packages). Persisted trace schema **1.0**. Requires **Node.js >= 20**.
201
+ </details>
157
202
 
158
- ```bash
159
- pnpm add agent-inspect
160
- npx agent-inspect doctor
161
- ```
203
+ ## Documentation
162
204
 
163
- Monorepo development: `pnpm install && pnpm build && pnpm test`
205
+ | | Website | Repo |
206
+ | - | ------- | ---- |
207
+ | Getting started | [docs](https://agentinspect.vercel.app/docs/getting-started/) | [FIRST-TRACE](https://github.com/rajudandigam/agent-inspect/blob/main/docs/FIRST-TRACE-IN-5-MINUTES.md) |
208
+ | Contracts / suites / gates | — | [TRACE-CONTRACTS](https://github.com/rajudandigam/agent-inspect/blob/main/docs/TRACE-CONTRACTS.md) · [SUITES-COHORTS-GATES](https://github.com/rajudandigam/agent-inspect/blob/main/docs/SUITES-COHORTS-GATES.md) |
209
+ | Safe sharing | [safe-sharing](https://agentinspect.vercel.app/docs/safe-sharing/) | [SAFE-TRACE-SHARING](https://github.com/rajudandigam/agent-inspect/blob/main/docs/SAFE-TRACE-SHARING.md) |
210
+ | API / CLI | — | [API](docs/API.md) · [CLI](docs/CLI.md) (packed with npm) |
211
+ | Full index | — | [docs/README.md](https://github.com/rajudandigam/agent-inspect/blob/main/docs/README.md) |
164
212
 
165
213
  ## Contributing
166
214
 
167
- [CONTRIBUTING.md](CONTRIBUTING.md) · [Good first issues](GOOD-FIRST-ISSUES.md) · [Discussions](https://github.com/rajudandigam/agent-inspect/discussions) · [Changelog](CHANGELOG.md)
215
+ [CONTRIBUTING.md](https://github.com/rajudandigam/agent-inspect/blob/main/CONTRIBUTING.md) · [Good first issues](https://github.com/rajudandigam/agent-inspect/blob/main/GOOD-FIRST-ISSUES.md) · [Discussions](https://github.com/rajudandigam/agent-inspect/discussions)
168
216
 
169
217
  **Redact traces before posting issues or PRs.**
218
+
219
+ ```bash
220
+ pnpm add agent-inspect
221
+ npx agent-inspect doctor
222
+ ```
223
+
224
+ Monorepo: `pnpm install && pnpm build && pnpm test`
package/docs/ADOPTION.md CHANGED
@@ -20,7 +20,7 @@ npx agent-inspect init --yes
20
20
  node examples/agent-inspect-demo.mjs
21
21
  npx agent-inspect list --dir .agent-inspect
22
22
  npx agent-inspect view <run-id> --dir .agent-inspect
23
- npx agent-inspect verify-safe --dir .agent-inspect
23
+ npx agent-inspect verify-safe <run-id> --dir .agent-inspect
24
24
  ```
25
25
 
26
26
  No API keys required for the init demo. See [examples/starters/](../examples/starters/README.md).
@@ -0,0 +1,67 @@
1
+ # Shareable trace bundles
2
+
3
+ AgentInspect **bundles** are local, offline folders you can attach to PRs, incident threads, or internal reviews. They combine redacted trace copies, HTML reports, safety check results, and a human summary — without mutating source traces or calling the network.
4
+
5
+ ## Quick start
6
+
7
+ ```bash
8
+ # Single run (default share profile + automatic verify-safe)
9
+ npx agent-inspect bundle <runId> --dir ./.agent-inspect
10
+
11
+ # Session or time window
12
+ npx agent-inspect bundle --session <sessionId> --dir ./.agent-inspect
13
+ npx agent-inspect bundle --since 24h --profile strict --dir ./.agent-inspect
14
+
15
+ # Explicit output folder (.zip suffix is stripped — folder-first MVP)
16
+ npx agent-inspect bundle <runId> --out ./my-bundle --json
17
+ ```
18
+
19
+ When a workspace exists, default output goes under `.agent-inspect/bundles/`.
20
+
21
+ ## Bundle layout
22
+
23
+ ```text
24
+ trace.html # offline HTML (index for multi-run)
25
+ trace.jsonl # redacted JSONL copy
26
+ summary.md # human overview
27
+ metadata.json # manifest (version, profile, safe status)
28
+ check-results.json # verify-safe results per run
29
+ redaction-report.json # detector summary (no secret values)
30
+ eval-results.json # placeholder unless eval artifacts are added later
31
+ performance-summary.json # placeholder unless perf artifacts are added later
32
+ assets/runs/<runId>.* # per-run HTML + JSONL mirrors
33
+ ```
34
+
35
+ ## Safety defaults
36
+
37
+ | Setting | Default |
38
+ | ------- | ------- |
39
+ | Redaction profile | `share` |
40
+ | verify-safe | Runs automatically before write |
41
+ | UNSAFE traces | Command fails unless `--allow-unsafe` |
42
+ | Source traces | Read-only; never modified |
43
+
44
+ Profiles:
45
+
46
+ - **`share`** — PR / internal support (default)
47
+ - **`strict`** — external or public sharing
48
+ - **`local`** — minimal redaction for local archives only
49
+
50
+ ## When to use bundles vs other commands
51
+
52
+ | Need | Command |
53
+ | ---- | ------- |
54
+ | One redacted file | `redact` |
55
+ | Safety scan only | `verify-safe` |
56
+ | CI artifact set | `artifacts` |
57
+ | PR-ready evidence folder | **`bundle`** |
58
+
59
+ ## Review before sharing
60
+
61
+ Bundles are **derived copies**, not compliance certification. Always review `summary.md`, `check-results.json`, and `trace.html` before attaching to tickets or PRs.
62
+
63
+ See also [SAFE-TRACE-SHARING.md](./SAFE-TRACE-SHARING.md) and [CLI.md §6.24](./CLI.md#624-bundle).
64
+
65
+ ## RFC
66
+
67
+ Design details: [proposals/SHAREABLE-BUNDLES-V4.3.md](./proposals/SHAREABLE-BUNDLES-V4.3.md).
package/docs/COMPARE.md CHANGED
@@ -10,11 +10,12 @@ AgentInspect is a local-first trace workbench for TypeScript AI agents: capture
10
10
  | --------- | ------------ | --------------------------------------- | --------------------------- | ------------------------------- |
11
11
  | Where traces live | Local JSONL on your disk | Vendor or self-hosted backend | Vendor backend | Your collector + backend |
12
12
  | Account / setup | None; `npm install` + CLI | Account or deployment | Account | SDK + collector + backend |
13
- | Primary surface | CLI (`view`, `report`, `diff`, `check`), local viewer, in-repo VS Code extension | Web dashboards | Web dashboards, eval UI | Dashboards, alerting |
13
+ | Primary surface | CLI (`view`, `report`, `diff`, `check`, suites/gates), local viewer, optional Studio Beta | Web dashboards | Web dashboards, eval UI | Dashboards, alerting |
14
14
  | Evals | Deterministic local heuristics and CI gates (`check`, `eval`, `circuit`, `guardrails`) | Platform evals | Datasets, scoring, regressions at scale | Not the focus |
15
15
  | Retention / fleet view | Not the goal | Strong | Strong | Strong |
16
16
  | Data sharing | Explicit `redact` + `scan` / `verify-safe` before you share a file | Team access controls | Team access controls | Org pipelines |
17
- | Network behavior | No upload; everything stays local unless you share a file | Uploads traces by design | Uploads by design | Ships telemetry by design |
17
+ | Network behavior | No default upload; explicit customer-owned ingest only | Uploads traces by design | Uploads by design | Ships telemetry by design |
18
+ | Team review | Optional customer-owned Studio Beta (not maintainer-hosted) | Strong | Strong | Strong |
18
19
 
19
20
  If you need hosted retention, fleet dashboards, dataset management, or org-wide pipelines, use one of those platforms alongside AgentInspect. Boundaries are listed in [LIMITATIONS.md](./LIMITATIONS.md); concrete inner-loop workflows in [USE-CASES.md](./USE-CASES.md).
20
21
 
@@ -42,19 +42,20 @@ Optional: `npx agent-inspect serve --dir .agent-inspect` for browser viewer.
42
42
  ## Beat 4 — Verify (30s)
43
43
 
44
44
  ```bash
45
- npx agent-inspect check .agent-inspect/*.jsonl --require-completed
45
+ npx agent-inspect check <run-id> --dir .agent-inspect
46
46
  ```
47
47
 
48
- Mention `--detect-stalls` and `--max-step-duration 30s` for long-running agents.
48
+ Optional flags: `--require-completed`, `--detect-stalls`, `--max-step-duration 30s`.
49
49
 
50
50
  ## Beat 5 — Share safely (30s)
51
51
 
52
52
  ```bash
53
- npx agent-inspect redact .agent-inspect/*.jsonl --profile share -o safe.jsonl
54
- npx agent-inspect verify-safe safe.jsonl
53
+ npx agent-inspect redact <run-id> --dir .agent-inspect --profile share -o safe.jsonl
54
+ npx agent-inspect verify-safe <run-id> --dir .agent-inspect
55
+ # or: npx agent-inspect verify-safe safe.jsonl
55
56
  ```
56
57
 
57
- "Redacted copy is safe to attach to a GitHub issue or Slack."
58
+ "Redacted / verified-safe artifacts are safe to attach to a GitHub issue or Slack."
58
59
 
59
60
  ## Beat 6 — Fix and diff (optional, 30s)
60
61
 
@@ -1,13 +1,23 @@
1
1
  # First trace in 5 minutes
2
2
 
3
- Goal: install → one trace → one check → one share-safe artifact.
3
+ Goal: install → one trace → one check → one share-safe bundle.
4
4
 
5
5
  **Docs site:** [https://agentinspect.vercel.app/docs/getting-started/](https://agentinspect.vercel.app/docs/getting-started/)
6
6
 
7
7
  ```bash
8
- npm install agent-inspect && npx agent-inspect init --yes && node examples/agent-inspect-demo.mjs
9
- npx agent-inspect list --dir .agent-inspect && npx agent-inspect report <run-id> --dir .agent-inspect
10
- npx agent-inspect check .agent-inspect/*.jsonl && npx agent-inspect verify-safe --dir .agent-inspect
8
+ npm install agent-inspect
9
+ npx agent-inspect init --yes
10
+ node examples/agent-inspect-demo.mjs
11
+ npx agent-inspect list --dir .agent-inspect
12
+ ```
13
+
14
+ Copy a `<run-id>` from `list`, then:
15
+
16
+ ```bash
17
+ npx agent-inspect report <run-id> --dir .agent-inspect
18
+ npx agent-inspect check <run-id> --dir .agent-inspect
19
+ npx agent-inspect bundle <run-id> --dir .agent-inspect --profile share
20
+ npx agent-inspect verify-safe <run-id> --dir .agent-inspect
11
21
  ```
12
22
 
13
23
  ## Minutes 0–1: Install
@@ -18,7 +28,8 @@ npm install agent-inspect
18
28
  npx agent-inspect init --yes
19
29
  ```
20
30
 
21
- Creates `agent-inspect.config.ts`, `.agent-inspect/`, and `examples/agent-inspect-demo.mjs`.
31
+ Creates `agent-inspect.config.ts`, `.agent-inspect/`, and `examples/agent-inspect-demo.mjs`.
32
+ `init` scaffolds files; it does **not** write a trace by itself.
22
33
 
23
34
  ## Minutes 1–2: Run
24
35
 
@@ -39,17 +50,23 @@ npx agent-inspect report <run-id> --dir .agent-inspect
39
50
  ## Minutes 3–4: Check
40
51
 
41
52
  ```bash
42
- npx agent-inspect check .agent-inspect/*.jsonl
53
+ npx agent-inspect check <run-id> --dir .agent-inspect
43
54
  ```
44
55
 
45
56
  ## Minutes 4–5: Share-safe artifact
46
57
 
47
58
  ```bash
48
- npx agent-inspect redact --profile share --dir .agent-inspect
49
- npx agent-inspect verify-safe --dir .agent-inspect
59
+ npx agent-inspect bundle <run-id> --dir .agent-inspect --profile share
60
+ npx agent-inspect verify-safe <run-id> --dir .agent-inspect
50
61
  ```
51
62
 
52
- Attach redacted JSONL to a PR or issue — not raw traces.
63
+ Attach the share-profile bundle (or a redacted file) to a PR or issue — not raw traces.
64
+
65
+ Optional file redaction:
66
+
67
+ ```bash
68
+ npx agent-inspect redact <run-id> --dir .agent-inspect --profile share -o redacted.jsonl
69
+ ```
53
70
 
54
71
  ## Next steps
55
72
 
@@ -58,7 +75,8 @@ Attach redacted JSONL to a PR or issue — not raw traces.
58
75
  | AI SDK | [AI SDK adoption](./AI-SDK-ADOPTION.md) |
59
76
  | OpenAI Agents | [OpenAI Agents local](./OPENAI-AGENTS-LOCAL.md) |
60
77
  | LangChain | [Adapters](./ADAPTERS.md) |
61
- | CI tests | [CI artifacts](./CI-ARTIFACTS.md) |
78
+ | CI tests | [CI artifacts](./CI-ARTIFACTS.md) · [Suites / gates](./SUITES-COHORTS-GATES.md) |
79
+ | Golden path | [GOLDEN-PATH.md](./GOLDEN-PATH.md) |
62
80
  | Broken agent demo | [broken-agent-debugging starter](../examples/starters/broken-agent-debugging/README.md) |
63
81
 
64
82
  Full index: [docs/README.md](./README.md) · Website docs: [agentinspect.vercel.app/docs](https://agentinspect.vercel.app/docs/)
@@ -0,0 +1,40 @@
1
+ # Golden path (technical launch candidate)
2
+
3
+ The 6.7.x technical launch candidate path. Be precise about what is automated today.
4
+
5
+ ## Automated packed path (CI)
6
+
7
+ From a packed `agent-inspect` install (`scripts/packed-quickstart-e2e.mjs` via `pnpm pack:smoke`):
8
+
9
+ ```text
10
+ init --yes → demo → list → verify-safe <runId> --dir .agent-inspect
11
+ ```
12
+
13
+ This proves install, scaffold, one local run, and safety verification.
14
+
15
+ ## Recommended developer path (manual)
16
+
17
+ ```text
18
+ init → demo → list → view/report → check → bundle --profile share → verify-safe
19
+ ```
20
+
21
+ Use required positional targets (`<run-id>` or file path). See the root README five-minute path.
22
+
23
+ ## Technical-but-not-fully-automated
24
+
25
+ `scripts/golden-path-e2e.mjs` covers a subset (init → inspect → verify-safe → report). It does **not** yet automate the full roadmap chain:
26
+
27
+ ```text
28
+ broken → contract fail → suite → CI gate → fix → diff/cohort → Studio review
29
+ ```
30
+
31
+ ## External pilot step
32
+
33
+ Three external teams + design-partner Studio trial remain **pending**. Do not mark launch adoption complete without rows in [implementation/PRE-V7-ADOPTION-EVIDENCE.md](./implementation/PRE-V7-ADOPTION-EVIDENCE.md). Pilot kit: [PRE-V7-PILOT-KIT.md](./PRE-V7-PILOT-KIT.md).
34
+
35
+ ## Pending product gaps (honest)
36
+
37
+ - TraceContract matchers / full contract depth
38
+ - Studio HTML productization for some pages
39
+ - Cross-platform consumer matrix evidence
40
+ - External Collector/Phoenix verification