tavus-cli 0.10.7__tar.gz → 0.12.0__tar.gz

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.
Files changed (132) hide show
  1. tavus_cli-0.12.0/.cursor/BUGBOT.md +33 -0
  2. tavus_cli-0.12.0/.github/review-guidelines.md +37 -0
  3. tavus_cli-0.12.0/.github/workflows/claude-pr-review.yml +68 -0
  4. tavus_cli-0.12.0/.github/workflows/codex-pr-review.yml +303 -0
  5. tavus_cli-0.12.0/.release-please-manifest.json +3 -0
  6. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/CHANGELOG.md +20 -0
  7. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/PKG-INFO +21 -2
  8. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/PROMPT.md +2 -2
  9. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/README.md +20 -1
  10. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/docs/chatgpt-plugin.md +7 -6
  11. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/docs/gemini-enterprise/README.md +2 -2
  12. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/docs/gemini-enterprise/test-queries.md +10 -11
  13. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/packages/tavus/pyproject.toml +2 -2
  14. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/packages/tavus-mcp/pyproject.toml +2 -2
  15. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/plugins/tavus-pal/plugin.json +7 -7
  16. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/pyproject.toml +1 -1
  17. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/skills/tavus-agentic-pals/SKILL.md +26 -5
  18. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/skills/tavus-agentic-pals/references/builder-workflow.md +3 -2
  19. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/__init__.py +1 -1
  20. tavus_cli-0.12.0/src/tavus_mcp/builder_guidance.json +62 -0
  21. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/cli/main.py +8 -22
  22. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/client/http.py +41 -2
  23. tavus_cli-0.12.0/src/tavus_mcp/sdk/pal_urls.py +53 -0
  24. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/build_and_verify.py +2 -6
  25. tavus_cli-0.12.0/src/tavus_mcp/sdk/recipes/conversations.py +70 -0
  26. tavus_cli-0.12.0/src/tavus_mcp/sdk/recipes/face_studio.py +208 -0
  27. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/quickstart.py +2 -12
  28. tavus_cli-0.12.0/src/tavus_mcp/sdk/schemas/face_studio.py +63 -0
  29. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/server.py +187 -49
  30. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/integration/test_client.py +4 -6
  31. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/integration/test_mcp_server.py +248 -3
  32. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_build_and_verify.py +2 -0
  33. tavus_cli-0.12.0/tests/unit/test_conversations.py +56 -0
  34. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/uv.lock +1 -1
  35. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/package-lock.json +2 -2
  36. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/package.json +1 -1
  37. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/src/index.ts +18 -3
  38. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/src/oauth-grant.ts +1 -1
  39. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/src/oauth.ts +24 -18
  40. tavus_cli-0.12.0/workers/tavus-mcp/src/slack-link-store.ts +40 -0
  41. tavus_cli-0.12.0/workers/tavus-mcp/src/slack.ts +236 -0
  42. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/src/tavus.ts +406 -68
  43. tavus_cli-0.12.0/workers/tavus-mcp/src/version.ts +1 -0
  44. tavus_cli-0.12.0/workers/tavus-mcp/test/conversations.test.ts +73 -0
  45. tavus_cli-0.12.0/workers/tavus-mcp/test/slack-link.test.ts +268 -0
  46. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/test/tavus.test.ts +240 -6
  47. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/test/worker-runtime.ts +11 -6
  48. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/wrangler.jsonc +10 -5
  49. tavus_cli-0.10.7/.release-please-manifest.json +0 -3
  50. tavus_cli-0.10.7/src/tavus_mcp/builder_guidance.json +0 -26
  51. tavus_cli-0.10.7/src/tavus_mcp/sdk/pal_urls.py +0 -24
  52. tavus_cli-0.10.7/workers/tavus-mcp/src/version.ts +0 -1
  53. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/.agents/plugins/marketplace.json +0 -0
  54. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/.env.example +0 -0
  55. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/.github/workflows/ci.yml +0 -0
  56. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/.github/workflows/deploy-cloudflare-worker.yml +0 -0
  57. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/.github/workflows/pr-reviews.yml +0 -0
  58. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/.github/workflows/pypi-cli.yml +0 -0
  59. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/.github/workflows/tail-cloudflare-worker.yml +0 -0
  60. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/.github/workflows/validate-staging-deployment.yml +0 -0
  61. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/.gitignore +0 -0
  62. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/AGENTS.md +0 -0
  63. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/docs/builder-guidance.md +0 -0
  64. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/docs/gemini-enterprise/listing.md +0 -0
  65. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/docs/usage-observability.md +0 -0
  66. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/packages/tavus/README.md +0 -0
  67. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/packages/tavus-mcp/README.md +0 -0
  68. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/plugins/tavus-pal/assets/icon-dark.svg +0 -0
  69. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/plugins/tavus-pal/assets/icon.svg +0 -0
  70. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/plugins/tavus-pal/mcp.json +0 -0
  71. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/release-please-config.json +0 -0
  72. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/scripts/__init__.py +0 -0
  73. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/scripts/build_chatgpt_plugin.py +0 -0
  74. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/scripts/check_release_versions.py +0 -0
  75. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/skills/tavus-agentic-pals/agents/openai.yaml +0 -0
  76. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/skills/tavus-agentic-pals/references/build-and-verify.md +0 -0
  77. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/skills/tavus-agentic-pals/references/chat-mode.md +0 -0
  78. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/skills/tavus-agentic-pals/references/examples.md +0 -0
  79. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/builder_guidance.py +0 -0
  80. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/cli/__init__.py +0 -0
  81. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/__init__.py +0 -0
  82. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/auth/__init__.py +0 -0
  83. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/auth/keyring_store.py +0 -0
  84. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/auth/oauth.py +0 -0
  85. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/auth/session.py +0 -0
  86. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/client/__init__.py +0 -0
  87. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/daily_transport.py +0 -0
  88. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/env.py +0 -0
  89. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/errors.py +0 -0
  90. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/patch.py +0 -0
  91. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/__init__.py +0 -0
  92. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/capabilities.py +0 -0
  93. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/carrier_builder.py +0 -0
  94. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/deployments.py +0 -0
  95. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/knowledge.py +0 -0
  96. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/meetings.py +0 -0
  97. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/options.py +0 -0
  98. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/scaffold_embed.py +0 -0
  99. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/templates.py +0 -0
  100. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/recipes/tool_reuse.py +0 -0
  101. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/schemas/__init__.py +0 -0
  102. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/schemas/deployment.py +0 -0
  103. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/schemas/file_manifest.py +0 -0
  104. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/schemas/guardrail.py +0 -0
  105. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/schemas/objective.py +0 -0
  106. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/schemas/pal.py +0 -0
  107. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/schemas/pronunciation.py +0 -0
  108. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/src/tavus_mcp/sdk/schemas/tool.py +0 -0
  109. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/conftest.py +0 -0
  110. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/integration/test_cli.py +0 -0
  111. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/integration/test_oauth.py +0 -0
  112. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_auth.py +0 -0
  113. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_builder.py +0 -0
  114. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_capabilities.py +0 -0
  115. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_daily_transport.py +0 -0
  116. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_deployments.py +0 -0
  117. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_env.py +0 -0
  118. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_guardrail_schema.py +0 -0
  119. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_objective_schema.py +0 -0
  120. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_pal_schema.py +0 -0
  121. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_patch.py +0 -0
  122. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_pronunciation_schema.py +0 -0
  123. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_release_versions.py +0 -0
  124. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_scaffold_embed.py +0 -0
  125. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_tool_reuse.py +0 -0
  126. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/tests/unit/test_tool_schema.py +0 -0
  127. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/.dev.vars.example +0 -0
  128. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/src/oauth-session.ts +0 -0
  129. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/test/builder-guidance.test.ts +0 -0
  130. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/test/oauth-session.test.ts +0 -0
  131. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/test/parity.test.ts +0 -0
  132. {tavus_cli-0.10.7 → tavus_cli-0.12.0}/workers/tavus-mcp/tsconfig.json +0 -0
@@ -0,0 +1,33 @@
1
+ # Bugbot review guidelines
2
+
3
+ ## Writing style
4
+
5
+ Write the review summary and every comment in plain language (ISO 24495-1),
6
+ matching the standard this repo applies to all agent-written output (see the
7
+ "Plain language for published writing" section of AGENTS.md):
8
+
9
+ - Put the most important point first.
10
+ - Use short sentences, active voice, and everyday words.
11
+ - Expand project shorthand (RQH, CVI, PAL, carrier) on first use.
12
+ - End each finding with a concrete next step for the author: what to change,
13
+ decide, or verify.
14
+
15
+ ## Credentials and identity
16
+
17
+ The Cloudflare Worker in `workers/tavus-mcp/` holds customers' Tavus API keys.
18
+ Treat these as high severity:
19
+
20
+ - A secret, API key, token, or code that can be forwarded to the Tavus API or
21
+ another service, logged, or returned to a caller that should not see it,
22
+ including on fallback paths such as a missing header or an unset secret.
23
+ - A code, state, or link that is not bound to the client, user, and flow that
24
+ issued it, or whose expiry or one-time use is not enforced.
25
+ - A claim in the PR description or a code comment ("works once", "expires",
26
+ "only the bot can") that the code does not enforce.
27
+
28
+ ## Tools and tests
29
+
30
+ - Flag a tool whose Python version (`src/tavus_mcp/`) and Worker version
31
+ (`workers/tavus-mcp/src/tavus.ts`) differ in parameters, defaults, or output.
32
+ - Flag a fix without a test that fails on the previous code, and a security
33
+ check without a test for each way it rejects a request.
@@ -0,0 +1,37 @@
1
+ Business context:
2
+
3
+ - This repo is the Tavus MCP server. Two implementations expose the same tools: the Python package (`src/tavus_mcp/`, the local server and `tavus-cli` on PyPI) and the Cloudflare Worker (`workers/tavus-mcp/`) that serves the hosted server at mcp.tavus.io. `workers/tavus-mcp/test/parity.test.ts` keeps their tool lists in step.
4
+ - Agents such as Claude, ChatGPT, Cursor and a Slack bot use these tools to build and manage a creator's PALs (Tavus conversational video agents) with that creator's own Tavus account.
5
+ - The Worker holds customers' Tavus API keys: OAuth sessions, sign-in state, and linked accounts. A bug there can expose a customer's key or let one caller act as another account.
6
+
7
+ Review objective:
8
+
9
+ - Perform a general code review for correctness, security, maintainability, and tests across the full PR.
10
+ - Treat a leaked or misrouted credential, one caller acting as another account, and runtime crashes as high severity.
11
+
12
+ Focus areas (in order):
13
+
14
+ 1. Credentials and identity in the Worker
15
+
16
+ - Every secret, API key, token, code, or signed state: where it can travel. Flag any path that forwards it to the Tavus API or another service, logs it, puts a plaintext key in a URL, or returns it to a caller that should not see it. Check the fallback paths too, such as a missing header, an unset secret, or an unrecognized bearer.
17
+ - Every code, state, or link a flow issues: whether it is bound to the client, user, and flow that issued it, whether its expiry is checked, and whether it can be swapped for one issued to someone else. OAuth uses PKCE for this binding; a new flow needs an equivalent.
18
+ - Claims in the PR description or code comments, such as "works once", "expires", or "only the bot can", that the code does not enforce.
19
+ - Secret comparisons that are not constant-time, and secrets the code does not check for presence and minimum length.
20
+ - New Durable Object classes need a migration in `wrangler.jsonc` and a binding in every environment.
21
+
22
+ 2. Tool correctness
23
+
24
+ - The Python and Worker versions of a tool accept the same parameters, apply the same defaults, and return the same shape.
25
+ - Parameters sent to the Tavus API use values the API accepts.
26
+ - Tool annotations (`readOnlyHint`, `destructiveHint`) match what the tool does.
27
+ - Tools hand off PALs with PAL Maker links. They never return a conversation's call link (see `PROMPT.md`).
28
+
29
+ 3. Tests
30
+
31
+ - A fix needs a test that fails on the previous code.
32
+ - A security check needs a test for each way it rejects a request, not only the success path.
33
+ - Test names match what the test asserts.
34
+
35
+ 4. Published text
36
+
37
+ - MCP tool descriptions, CLI `--help` text, the README, and PR descriptions follow the plain language rules in `AGENTS.md`. Tool descriptions ship to end users.
@@ -0,0 +1,68 @@
1
+ name: Claude PR Review
2
+
3
+ on:
4
+ pull_request:
5
+ types: [opened, reopened, ready_for_review]
6
+ branches: [main]
7
+
8
+ permissions:
9
+ contents: read
10
+ issues: write
11
+ pull-requests: write
12
+ # Required by claude-code-action@v1, which fetches an OIDC token even when
13
+ # authenticating with an API key. Without it the action aborts.
14
+ id-token: write
15
+
16
+ jobs:
17
+ claude-review:
18
+ # Fork pull requests get no secrets, so skip them instead of failing.
19
+ if: github.event.pull_request.draft == false && !startsWith(github.head_ref, 'release-please--') && github.event.pull_request.head.repo.full_name == github.repository
20
+ runs-on: ubuntu-latest
21
+ env:
22
+ # One place for the model: it is passed to --model and named in the
23
+ # signature on every comment, so a stale model shows up in the PR.
24
+ REVIEW_MODEL: claude-opus-5-5
25
+
26
+ steps:
27
+ - name: Checkout repository
28
+ uses: actions/checkout@v5
29
+ with:
30
+ fetch-depth: 1
31
+ persist-credentials: false
32
+
33
+ - name: Load review guidelines
34
+ id: guidelines
35
+ # The Claude and Codex reviews share this text; edit it there, not in
36
+ # either prompt. Bugbot reads .cursor/BUGBOT.md instead.
37
+ run: |
38
+ {
39
+ echo 'text<<GUIDELINES_EOF'
40
+ cat .github/review-guidelines.md
41
+ echo 'GUIDELINES_EOF'
42
+ } >> "$GITHUB_OUTPUT"
43
+
44
+ - name: Claude code review
45
+ id: claude_review
46
+ uses: anthropics/claude-code-action@1d6de8cb0c237e7c15e9e1bdf973826ebae490cc # v1.0.248
47
+ with:
48
+ anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
49
+ prompt: |
50
+ REPO: ${{ github.repository }}
51
+ PR NUMBER: ${{ github.event.pull_request.number }}
52
+ You are reviewing the Tavus MCP server: a Python package and a Cloudflare Worker that expose the same tools.
53
+
54
+ ${{ steps.guidelines.outputs.text }}
55
+
56
+ Output requirements:
57
+ - Leave inline comments for concrete issues.
58
+ - Do not repeat already-reported inline issues unless the code changed or there is new evidence.
59
+ - Provide a final summary grouped by severity: Critical, High, Medium, Low.
60
+ - For each Critical/High issue include:
61
+ - why it matters (user impact)
62
+ - a specific fix suggestion
63
+ - If no major risks are found, explicitly state why this PR is low risk.
64
+ - End the final summary and every inline comment with this line on its own: `Reviewed by ${{ env.REVIEW_MODEL }}`
65
+ - Write the summary and every comment in plain language (ISO 24495-1): most important point first, short sentences, active voice, project shorthand expanded on first use, and a concrete next step for the author.
66
+ claude_args: |
67
+ --model ${{ env.REVIEW_MODEL }}
68
+ --allowedTools "mcp__github_inline_comment__create_inline_comment,Bash(gh pr comment:*),Bash(gh pr diff:*),Bash(gh pr view:*)"
@@ -0,0 +1,303 @@
1
+ name: Codex PR Review
2
+
3
+ on:
4
+ # Automatic review for public PRs: non-draft PRs run on open/reopen,
5
+ # and draft PRs run once they are marked ready for review.
6
+ pull_request:
7
+ types: [opened, reopened, ready_for_review]
8
+ branches: [main]
9
+
10
+ permissions:
11
+ contents: read
12
+
13
+ jobs:
14
+ codex-review:
15
+ # Fork pull requests get no secrets, so skip them instead of failing.
16
+ if: github.event.pull_request.draft == false && !startsWith(github.head_ref, 'release-please--') && github.event.pull_request.head.repo.full_name == github.repository
17
+ runs-on: ubuntu-latest
18
+ # Reviews normally finish in 3-9 minutes. Without a timeout, a hung Codex
19
+ # API call burns the 6-hour default; observed hangs died at ~64 minutes
20
+ # with "runner lost communication" instead of a clear failure.
21
+ timeout-minutes: 20
22
+ # The review job only reads the diff and emits structured output; the
23
+ # post_feedback job is the one that writes to the PR.
24
+ permissions:
25
+ contents: read
26
+ outputs:
27
+ final_message: ${{ steps.run_codex.outputs.final-message }}
28
+ reviewed_sha: ${{ steps.reviewed.outputs.sha }}
29
+
30
+ steps:
31
+ - name: Checkout repository
32
+ uses: actions/checkout@v5
33
+ with:
34
+ ref: refs/pull/${{ github.event.pull_request.number }}/merge
35
+ # Fetch the merge commit plus its parents so `git diff HEAD^1 HEAD`
36
+ # can resolve the base side of the diff.
37
+ fetch-depth: 2
38
+
39
+ - name: Record the reviewed head
40
+ id: reviewed
41
+ # The merge commit's second parent is the pull request head under review.
42
+ run: echo "sha=$(git rev-parse HEAD^2)" >> "$GITHUB_OUTPUT"
43
+
44
+ - name: Pre-fetch base and head refs
45
+ env:
46
+ PR_BASE_REF: ${{ github.event.pull_request.base.ref }}
47
+ PR_NUMBER: ${{ github.event.pull_request.number }}
48
+ run: |
49
+ git fetch --no-tags origin \
50
+ "+refs/heads/$PR_BASE_REF:refs/remotes/origin/$PR_BASE_REF" \
51
+ "+refs/pull/$PR_NUMBER/head:refs/remotes/pull/$PR_NUMBER/head"
52
+
53
+ - name: Drop persisted git credentials
54
+ # Checkout stores the job token in .git/config; remove it before Codex reads the tree.
55
+ run: |
56
+ git config --local --unset-all http.https://github.com/.extraheader || true
57
+ if git config --local --get-regexp 'http\..*\.extraheader'; then
58
+ echo "::error::persisted git credentials still present after cleanup"
59
+ exit 1
60
+ fi
61
+
62
+ - name: Load review guidelines
63
+ id: guidelines
64
+ # The Claude and Codex reviews share this text; edit it there, not in
65
+ # either prompt. Bugbot reads .cursor/BUGBOT.md instead.
66
+ run: |
67
+ {
68
+ echo 'text<<GUIDELINES_EOF'
69
+ cat .github/review-guidelines.md
70
+ echo 'GUIDELINES_EOF'
71
+ } >> "$GITHUB_OUTPUT"
72
+
73
+ - name: Codex code review
74
+ id: run_codex
75
+ # Pinned to v1.11: the moving v1 tag picked up v1.12 on 2026-08-20,
76
+ # whose Linux runner-setup rewrite never exits after the review
77
+ # completes (openai/codex-action#150). Unpin once that is fixed.
78
+ uses: openai/codex-action@52fe01ec70a42f454c9d2ebd47598f9fd6893d56
79
+ with:
80
+ openai-api-key: ${{ secrets.OPENAI_API_KEY }}
81
+ model: gpt-6-astra
82
+ effort: xhigh
83
+ prompt: |
84
+ REPO: ${{ github.repository }}
85
+ PR NUMBER: ${{ github.event.pull_request.number }}
86
+ You are Codex running GPT-6 Astra with Extra High thinking.
87
+ You are reviewing the Tavus MCP server: a Python package and a Cloudflare Worker that expose the same tools.
88
+
89
+ ${{ steps.guidelines.outputs.text }}
90
+
91
+ Output requirements:
92
+ - The workspace is checked out to the PR merge commit. Review the PR diff with `git diff HEAD^1 HEAD`.
93
+ - Return JSON that matches the provided schema.
94
+ - Put the review summary in `summary`, grouped by severity: Critical, High, Medium, Low.
95
+ - Put concrete line-specific findings in `inline_comments`.
96
+ - Only create inline comments for issues that can be anchored to a line in the PR diff.
97
+ - Leave `inline_comments` empty if there are no concrete line-specific findings.
98
+ - For each Critical/High issue include:
99
+ - why it matters (user impact)
100
+ - a specific fix suggestion
101
+ - If no major risks are found, explicitly state why this PR is low risk.
102
+ - Write the summary and every comment in plain language (ISO 24495-1): most important point first, short sentences, active voice, project shorthand expanded on first use, and a concrete next step for the author.
103
+ output-schema: |
104
+ {
105
+ "type": "object",
106
+ "additionalProperties": false,
107
+ "required": ["summary", "inline_comments"],
108
+ "properties": {
109
+ "summary": {
110
+ "type": "string",
111
+ "description": "Markdown summary for the overall PR review."
112
+ },
113
+ "inline_comments": {
114
+ "type": "array",
115
+ "maxItems": 20,
116
+ "items": {
117
+ "type": "object",
118
+ "additionalProperties": false,
119
+ "required": ["path", "line", "severity", "body"],
120
+ "properties": {
121
+ "path": {
122
+ "type": "string",
123
+ "description": "Repository-relative file path from the PR diff."
124
+ },
125
+ "line": {
126
+ "type": "integer",
127
+ "description": "Right-side file line number to anchor the comment."
128
+ },
129
+ "severity": {
130
+ "type": "string",
131
+ "enum": ["Critical", "High", "Medium", "Low"]
132
+ },
133
+ "body": {
134
+ "type": "string",
135
+ "description": "Markdown body for this inline comment."
136
+ }
137
+ }
138
+ }
139
+ }
140
+ }
141
+ }
142
+ sandbox: read-only
143
+
144
+ post_feedback:
145
+ runs-on: ubuntu-latest
146
+ needs: codex-review
147
+ if: needs.codex-review.outputs.final_message != ''
148
+ permissions:
149
+ issues: write
150
+ pull-requests: write
151
+ steps:
152
+ - name: Post Codex feedback
153
+ uses: actions/github-script@v8
154
+ env:
155
+ CODEX_FINAL_MESSAGE: ${{ needs.codex-review.outputs.final_message }}
156
+ REVIEWED_SHA: ${{ needs.codex-review.outputs.reviewed_sha }}
157
+ with:
158
+ github-token: ${{ github.token }}
159
+ script: |
160
+ function parseCodexOutput(raw) {
161
+ const trimmed = raw.trim();
162
+ const withoutFence = trimmed
163
+ .replace(/^```(?:json)?\s*/i, '')
164
+ .replace(/\s*```$/i, '')
165
+ .trim();
166
+ return JSON.parse(withoutFence);
167
+ }
168
+
169
+ function diffRightLines(patch) {
170
+ const lines = new Set();
171
+ if (!patch) {
172
+ return lines;
173
+ }
174
+
175
+ let rightLine = 0;
176
+ for (const line of patch.split('\n')) {
177
+ const hunk = line.match(/^@@ -\d+(?:,\d+)? \+(\d+)(?:,\d+)? @@/);
178
+ if (hunk) {
179
+ rightLine = Number(hunk[1]);
180
+ continue;
181
+ }
182
+
183
+ if (line.startsWith('+')) {
184
+ lines.add(rightLine);
185
+ rightLine += 1;
186
+ } else if (line.startsWith('-')) {
187
+ continue;
188
+ } else if (line.startsWith('\\')) {
189
+ continue;
190
+ } else {
191
+ lines.add(rightLine);
192
+ rightLine += 1;
193
+ }
194
+ }
195
+
196
+ return lines;
197
+ }
198
+
199
+ let review;
200
+ try {
201
+ review = parseCodexOutput(process.env.CODEX_FINAL_MESSAGE);
202
+ } catch (error) {
203
+ await github.rest.pulls.createReview({
204
+ owner: context.repo.owner,
205
+ repo: context.repo.repo,
206
+ pull_number: context.payload.pull_request.number,
207
+ event: 'COMMENT',
208
+ body: [
209
+ '<!-- codex-pr-review -->',
210
+ '**Codex review: GPT-6 Astra, Extra High thinking**',
211
+ '',
212
+ 'Codex returned output that could not be parsed as structured review JSON, so posting the raw review output.',
213
+ '',
214
+ process.env.CODEX_FINAL_MESSAGE,
215
+ ].join('\n'),
216
+ });
217
+ return;
218
+ }
219
+
220
+ const files = await github.paginate(github.rest.pulls.listFiles, {
221
+ owner: context.repo.owner,
222
+ repo: context.repo.repo,
223
+ pull_number: context.payload.pull_request.number,
224
+ per_page: 100,
225
+ });
226
+
227
+ const diffLinesByPath = new Map(
228
+ files.map((file) => [file.filename, diffRightLines(file.patch)])
229
+ );
230
+
231
+ // A push during the review shifts the diff, so the reviewed line
232
+ // numbers no longer point at the code Codex read.
233
+ const { data: pull } = await github.rest.pulls.get({
234
+ owner: context.repo.owner,
235
+ repo: context.repo.repo,
236
+ pull_number: context.payload.pull_request.number,
237
+ });
238
+ const headMoved = pull.head.sha !== process.env.REVIEWED_SHA;
239
+
240
+ const skipped = [];
241
+ const seen = new Set();
242
+ const comments = [];
243
+
244
+ for (const item of review.inline_comments || []) {
245
+ const path = String(item.path || '');
246
+ const line = Number(item.line);
247
+ const bodyText = String(item.body || '').trim();
248
+ const rawSeverity = String(item.severity || 'Medium');
249
+ const severity = ['Critical', 'High', 'Medium', 'Low'].includes(rawSeverity)
250
+ ? rawSeverity
251
+ : 'Medium';
252
+ const key = `${path}:${line}:${bodyText}`;
253
+ const validLines = diffLinesByPath.get(path);
254
+
255
+ if (!path || !Number.isInteger(line) || !bodyText || seen.has(key)) {
256
+ continue;
257
+ }
258
+ seen.add(key);
259
+
260
+ if (headMoved || !validLines || !validLines.has(line)) {
261
+ skipped.push({ path, line, severity, body: bodyText });
262
+ continue;
263
+ }
264
+
265
+ comments.push({
266
+ path,
267
+ line,
268
+ side: 'RIGHT',
269
+ body: `**Codex: GPT-6 Astra, Extra High**\n\n**${severity}:** ${bodyText}`,
270
+ });
271
+ }
272
+
273
+ const bodyParts = [
274
+ '<!-- codex-pr-review -->',
275
+ '**Codex review: GPT-6 Astra, Extra High thinking**',
276
+ '',
277
+ String(review.summary || '').trim() || 'Codex completed the review without a summary.',
278
+ ];
279
+
280
+ if (skipped.length > 0) {
281
+ bodyParts.push(
282
+ '',
283
+ headMoved
284
+ ? `The PR changed during the review, so ${skipped.length} finding(s) about commit ${process.env.REVIEWED_SHA.slice(0, 7)} are listed here instead of inline:`
285
+ : `${skipped.length} finding(s) could not be anchored to the current PR diff and are preserved here:`,
286
+ ...skipped.map((item) => `- **${item.severity}** \`${item.path}:${item.line}\`: ${item.body.replace(/\s*\n\s*/g, ' ')}`)
287
+ );
288
+ }
289
+
290
+ const reviewRequest = {
291
+ owner: context.repo.owner,
292
+ repo: context.repo.repo,
293
+ pull_number: context.payload.pull_request.number,
294
+ commit_id: process.env.REVIEWED_SHA,
295
+ event: 'COMMENT',
296
+ body: bodyParts.join('\n'),
297
+ };
298
+
299
+ if (comments.length > 0) {
300
+ reviewRequest.comments = comments;
301
+ }
302
+
303
+ await github.rest.pulls.createReview(reviewRequest);
@@ -0,0 +1,3 @@
1
+ {
2
+ ".": "0.12.0"
3
+ }
@@ -1,5 +1,25 @@
1
1
  # Changelog
2
2
 
3
+ ## [0.12.0](https://github.com/Tavus-Engineering/tavus-mcp/compare/v0.11.0...v0.12.0) (2026-10-27)
4
+
5
+
6
+ ### Features
7
+
8
+ * create faces from images and Face Studio through the MCP ([#92](https://github.com/Tavus-Engineering/tavus-mcp/issues/92)) ([e543a72](https://github.com/Tavus-Engineering/tavus-mcp/commit/e543a724742c619f28e1cd717a622213d4676b19))
9
+ * **slack:** act as a linked Slack user, and read past conversations ([#91](https://github.com/Tavus-Engineering/tavus-mcp/issues/91)) ([8830883](https://github.com/Tavus-Engineering/tavus-mcp/commit/883088379e2ab69cfd5aef4de20feb3774225a5a))
10
+
11
+
12
+ ### Continuous Integration
13
+
14
+ * review pull requests with Claude and Codex, and add Bugbot rules ([#94](https://github.com/Tavus-Engineering/tavus-mcp/issues/94)) ([519d399](https://github.com/Tavus-Engineering/tavus-mcp/commit/519d399a70c99e03460f39090a9cfb0bad43bac0))
15
+
16
+ ## [0.11.0](https://github.com/Tavus-Engineering/tavus-mcp/compare/v0.10.7...v0.11.0) (2026-10-21)
17
+
18
+
19
+ ### Features
20
+
21
+ * hand off PALs with PAL Maker links instead of call links ([#88](https://github.com/Tavus-Engineering/tavus-mcp/issues/88)) ([1def07d](https://github.com/Tavus-Engineering/tavus-mcp/commit/1def07d69a6299e5944dffd22d4bba1f8b7cbf7d))
22
+
3
23
  ## [0.10.7](https://github.com/Tavus-Engineering/tavus-mcp/compare/v0.10.6...v0.10.7) (2026-10-20)
4
24
 
5
25
 
@@ -1,6 +1,6 @@
1
1
  Metadata-Version: 2.5
2
2
  Name: tavus-cli
3
- Version: 0.10.7
3
+ Version: 0.12.0
4
4
  Summary: Tavus CLI and MCP server for agentic PAL creation and editing.
5
5
  Requires-Python: >=3.11
6
6
  Requires-Dist: fastmcp>=2.0.0
@@ -80,6 +80,25 @@ off to the embed or widget prompt in tavus-examples
80
80
  (`examples/embed-prompt.md`, `examples/widget-prompt.md`) with the
81
81
  deployment id and the PAL's page tools already filled in.
82
82
 
83
+ ### Slack identity for a Tavus Slack bot
84
+
85
+ A trusted Slack bot can call the hosted server on behalf of each Slack user,
86
+ so one bot serves many people without a shared Tavus key.
87
+
88
+ - Set the Worker secret `SLACK_BOT_SECRET` (random, at least 32 characters).
89
+ The bot sends it as `Authorization: Bearer <secret>` together with
90
+ `X-Slack-Team-Id` and `X-Slack-User-Id` headers on `/mcp`. A request that names
91
+ a Slack user without the secret gets `401`.
92
+ - A user who has not linked an account gets `403` with the error
93
+ `slack_account_not_linked`.
94
+ - To link, the bot calls `POST /slack/link` with `{"team_id", "user_id"}` and the
95
+ same bearer secret. The response has a one-time `url` that is good for ten
96
+ minutes. The user opens it, signs in through the normal Tavus browser sign-in,
97
+ and the Worker stores their account key encrypted. `DELETE /slack/link` with the
98
+ same body removes the link.
99
+ - Linking uses the `SLACK_LINKS` Durable Object, so a first deploy needs the `v3`
100
+ migration in `wrangler.jsonc`.
101
+
83
102
  ### ChatGPT plugin
84
103
 
85
104
  `plugins/tavus-pal/` packages this same hosted server as **Tavus PAL** for
@@ -182,7 +201,7 @@ Command groups:
182
201
  | `tavus face`, `tavus resource`, `tavus skill` | account resource discovery |
183
202
  | `tavus conversation` | create or end CVI conversations |
184
203
  | `tavus deployment` | `list`, `get`, `create` — put a PAL on a landing page, an embed or a widget |
185
- | `tavus auth`, `tavus doctor`, `tavus quickstart`, `tavus embed` | login, environment checks, one-call PAL + conversation, embed scaffolding |
204
+ | `tavus auth`, `tavus doctor`, `tavus quickstart`, `tavus embed` | login, environment checks, one-call PAL with test and deploy links, embed scaffolding |
186
205
 
187
206
  Examples:
188
207
 
@@ -36,12 +36,12 @@ How to do the common things properly
36
36
 
37
37
  These are the contracts the server keeps. Exact parameters and response shapes are in the reference below; the order of operations is here, because that is what tends to get invented.
38
38
 
39
- - Building a new PAL. Use `tavus_pal_build_and_verify`; it is the standard path. You run the interview first: ask me plain questions, up to 12, about who the PAL talks to, what a good conversation achieves, the tone, what it must never say or do, what it should know (links to documents), and what it should do in my app. Ask only what my description leaves open. Pass my description as `prompt`, the questions and my answers as `interview`, and your summary of each part as `briefs`. The call takes one to three minutes. It drafts the personality, objectives, guardrails and greeting, picks a face unless I asked for one (`face_id`), publishes the PAL, runs test conversations and judges them. Read `verdict.overall` (`pass`, `partial` or `fail`), then the summary and any failing evidence, before you tell me it works. Tools can only be attached once the PAL exists, so the build's test turns run without them and tool checks in the verdict fail; attach the tools, then rerun those turns with chat mode before you tell me the tools work. Give me the `pal_url` it returns; that opens the PAL in the PAL Maker. Use `tavus_pal_create` or `tavus_quickstart` only when I hand you a finished system prompt and ask you to skip the build.
39
+ - Building a new PAL. Use `tavus_pal_build_and_verify`; it is the standard path. You run the interview first: ask me plain questions, up to 12, about who the PAL talks to, what a good conversation achieves, the tone, what it must never say or do, what it should know (links to documents), and what it should do in my app. Ask only what my description leaves open. Pass my description as `prompt`, the questions and my answers as `interview`, and your summary of each part as `briefs`. The call takes one to three minutes. It drafts the personality, objectives, guardrails and greeting, picks a face unless I asked for one (`face_id`), publishes the PAL, runs test conversations and judges them. Read `verdict.overall` (`pass`, `partial` or `fail`), then the summary and any failing evidence, before you tell me it works. Tools can only be attached once the PAL exists, so the build's test turns run without them and tool checks in the verdict fail; attach the tools, then rerun those turns with chat mode before you tell me the tools work. Give me the `test_url` it returns, which opens the PAL in the PAL Maker and starts a video call with it, and the `pal_url`, which opens its prompt and settings. Then ask whether I want to deploy it, and describe each entry in `deployment_options`. Use `tavus_pal_create` or `tavus_quickstart` only when I hand you a finished system prompt and ask you to skip the build.
40
40
  - Changing an existing PAL. For a change to what the PAL does or shows, such as a new step, an on-screen scorecard or quiz, web search, or memory, call `tavus_pal_build_and_verify` with its `pal_id`, my request as `prompt`, and only the `briefs` for the parts that should change. Put built-in capabilities in `briefs.skills`. The same PAL Maker drafters edit the PAL and leave the rest alone, then it publishes and tests the PAL. Read `drafted` and `builtin_changes` in the result to see what changed. For an exact edit, such as a new greeting, call `tavus_pal_get` and `tavus_describe_pal_options` first, and take ids and patch paths from what they return. Make the smallest change with `tavus_patch_pal` (JSON Patch operations such as `replace` on `/greeting`), or use the dedicated tools for capabilities, Knowledge and tools. A change applies to every conversation the PAL starts from then on, including on my live site, so tell me what will change and get a yes before changing a PAL that is already in use.
41
41
  - Giving the PAL a tool. Check `tavus_tool_list` first and attach an existing tool when its name, parameters and delivery fit. When you create a new one that overlaps an existing tool, tell me which tool you saw and why it does not fit. The tool's `description` is what tells the PAL when to call it, so write it as an instruction. `delivery` decides where the call goes: `{ "app_message": true }` (the default) sends it to the page running the conversation, and `delivery.api` makes Tavus call a backend URL. Then write the handler in this repo with the exact tool name and parameter schema, so the two halves match. On a page that uses `<tavus-embed>` or `<tavus-widget>`, the call arrives as `tavus:tool-call` `{ name, arguments, tool_call_id }` with `arguments` as a JSON string, and the page replies with `sendMessage({ event_type: "conversation.tool_result", properties: { tool_call_id, output, status: "success" } })`; the host communication page below has the full contract. A tool that is not attached to the PAL never fires, however the code is wired.
42
42
  - Giving the PAL knowledge. Create a Knowledge document from a URL anyone can reach with `tavus_document_create`, then attach it with `tavus_pal_knowledge_add`. The hosted server cannot read files on my machine; for a local file, ask me to upload it in the PAL Maker's Knowledge section or to host it at a URL. Before you tell me what a document says, read what Tavus indexed with `tavus_document_chunks`.
43
43
  - Testing by text. `tavus_chat_start` starts a text-only conversation with the PAL, `tavus_chat_turn` sends one message and waits for the reply, and `tavus_chat_end` ends it. Always end the conversations you start. The first turn is slower while the conversation starts up. Replies are not deterministic, so judge behaviour (did it ask for the order number, did it refuse the pricing promise) rather than exact wording.
44
- - Seeing it on video. `tavus_pal_preview` starts a full video conversation and returns its URL for me to open. Do it after the text tests pass, and only when I want to see or hear the PAL.
44
+ - Seeing it on video. `tavus_pal_preview` returns the same PAL Maker links and deployment options for an existing PAL. It does not start a call. Never give me a conversation's call link; it opens a bare call page without the PAL Maker.
45
45
  - Putting the PAL in front of people. A deployment is what makes a PAL reachable, and there are three kinds. A landing page is a page Tavus hosts, with its own link, for when there is no site in this repo or I want something to send around. Embed puts the conversation inside a page of my app: a section, a card, a step in a flow. Widget is a floating launcher on every page of my site. Pick one from what this repo is, propose it with a reason, and let me decide. If I want the PAL to act on the page (open a product, fill a form), create and attach those tools before deploying, so the handoff below includes them.
46
46
  - Creating the deployment. Call `tavus_deployment_list` with the PAL id first and offer to reuse a deployment that already fits. Otherwise ask me what `tavus_deployment_create` needs: whether each visitor gets their own memory of past calls; for embed or widget, where the site is hosted, as origins such as `https://www.example.com`, plus `http://localhost:<port>` from this repo's dev server if I want to try it locally (an empty list lets any site load it); for a landing page, a headline and one sentence you draft from what the PAL does; and how much use to allow, since every call is billed and a deployment has no limits unless I set them: calls per day across all visitors, calls per day per visitor, calls at the same moment, and the longest a call may last. Suggest numbers that fit what I am building, and let me say no limit. A deployment is live and public as soon as it is created, so show me these values and get a yes first.
47
47
  - Handing off. `tavus_deployment_create` returns an `integration` block. For a landing page, give me `landing_page_url`. For embed or widget, read `prompt_url` and follow that prompt with `prompt_header` placed above it, so it does not ask for the deployment id or the page tools again; write a handler for every entry in `page_tools`, reusing any handler you already wrote in this session. When the PAL's tools change later, `tavus_deployment_get` returns the current list. Changing a deployment's settings or taking it down happens in the [Tavus PAL Maker](https://maker.tavus.io). Do not use `tavus_scaffold_embed` for a site: it wraps the URL of a single conversation, which stops working when that conversation ends.
@@ -58,6 +58,25 @@ off to the embed or widget prompt in tavus-examples
58
58
  (`examples/embed-prompt.md`, `examples/widget-prompt.md`) with the
59
59
  deployment id and the PAL's page tools already filled in.
60
60
 
61
+ ### Slack identity for a Tavus Slack bot
62
+
63
+ A trusted Slack bot can call the hosted server on behalf of each Slack user,
64
+ so one bot serves many people without a shared Tavus key.
65
+
66
+ - Set the Worker secret `SLACK_BOT_SECRET` (random, at least 32 characters).
67
+ The bot sends it as `Authorization: Bearer <secret>` together with
68
+ `X-Slack-Team-Id` and `X-Slack-User-Id` headers on `/mcp`. A request that names
69
+ a Slack user without the secret gets `401`.
70
+ - A user who has not linked an account gets `403` with the error
71
+ `slack_account_not_linked`.
72
+ - To link, the bot calls `POST /slack/link` with `{"team_id", "user_id"}` and the
73
+ same bearer secret. The response has a one-time `url` that is good for ten
74
+ minutes. The user opens it, signs in through the normal Tavus browser sign-in,
75
+ and the Worker stores their account key encrypted. `DELETE /slack/link` with the
76
+ same body removes the link.
77
+ - Linking uses the `SLACK_LINKS` Durable Object, so a first deploy needs the `v3`
78
+ migration in `wrangler.jsonc`.
79
+
61
80
  ### ChatGPT plugin
62
81
 
63
82
  `plugins/tavus-pal/` packages this same hosted server as **Tavus PAL** for
@@ -160,7 +179,7 @@ Command groups:
160
179
  | `tavus face`, `tavus resource`, `tavus skill` | account resource discovery |
161
180
  | `tavus conversation` | create or end CVI conversations |
162
181
  | `tavus deployment` | `list`, `get`, `create` — put a PAL on a landing page, an embed or a widget |
163
- | `tavus auth`, `tavus doctor`, `tavus quickstart`, `tavus embed` | login, environment checks, one-call PAL + conversation, embed scaffolding |
182
+ | `tavus auth`, `tavus doctor`, `tavus quickstart`, `tavus embed` | login, environment checks, one-call PAL with test and deploy links, embed scaffolding |
164
183
 
165
184
  Examples:
166
185
 
@@ -11,7 +11,7 @@ second server or copied tool registry to maintain.
11
11
  python scripts/build_chatgpt_plugin.py
12
12
  ```
13
13
 
14
- This produces `dist/tavus-pal-0.1.0.zip`. The ZIP contains only `plugin.json`,
14
+ This produces `dist/tavus-pal-0.1.2.zip`. The ZIP contains only `plugin.json`,
15
15
  `mcp.json`, and two square SVG icons derived from the portal's Tavus favicon.
16
16
  No account credentials, local environment files, or test evidence are bundled.
17
17
  The portable Agent Plugins schemas validate both manifests. The package has
@@ -42,8 +42,8 @@ listing when appropriate rather than creating a duplicate.
42
42
 
43
43
  The current hosted server exposes 67 tools. It can create and edit PALs,
44
44
  choose available faces and voices, attach knowledge and capabilities, configure
45
- guardrails and objectives, run text conversations, and return a browser video
46
- preview URL. The guided build tool drafts and publishes a PAL and evaluates
45
+ guardrails and objectives, run text conversations, return PAL Maker links to
46
+ test a PAL on video, and create deployments. The guided build tool drafts and publishes a PAL and evaluates
47
47
  conversation probes. Creation, building, and conversations use the connected
48
48
  Tavus account's existing entitlements and may consume Tavus usage.
49
49
 
@@ -53,8 +53,9 @@ PAL marketplace listings, purchase, or ownership transfer. Hosted MCP also
53
53
  cannot read local file paths; Python-only upload and builder-administration
54
54
  tools remain the documented parity exceptions.
55
55
 
56
- The package does not add an embedded video widget. Video preview opens the
57
- Tavus conversation URL, matching existing MCP behavior. OpenAI supports optional
56
+ The package does not add an embedded video widget. Video preview opens the PAL
57
+ in the Tavus PAL Maker's test mode; the MCP never returns a conversation's call
58
+ link. OpenAI supports optional
58
59
  MCP UI, but microphone/camera behavior in an embedded Tavus call still needs
59
60
  separate implementation and host testing before promising an in-chat video call.
60
61
 
@@ -153,7 +154,7 @@ not a claim that the whole review suite has passed.
153
154
  | Help me build a museum guide that explains paintings briefly. | `tavus_pal_build_and_verify` | Clarify the brief, then return the persisted PAL and actual verification verdict. |
154
155
  | Rename that PAL to Museum Guide. | `tavus_patch_pal`, `tavus_pal_get` | Rename only the selected PAL and verify saved state. |
155
156
  | Test my guide by asking what Impressionism is, then end the test. | `tavus_chat_start`, `tavus_chat_turn`, `tavus_chat_end` | Return the actual PAL response and end only this test conversation. |
156
- | Open a video preview of that guide. | `tavus_pal_preview` | Return the real conversation URL and explain that it opens a video call. |
157
+ | Open a video preview of that guide. | `tavus_pal_preview` | Return the PAL Maker test link, then offer the deployment options. No call link. |
157
158
 
158
159
  | Negative prompt | Expected behavior |
159
160
  | --- | --- |
@@ -104,9 +104,9 @@ forwarded from the bearer token.
104
104
  | `tavus_pal_from_template` | `POST /v2/pals` |
105
105
  | `tavus_pal_delete` | `DELETE /v2/pals/{id}` |
106
106
  | `tavus_patch_pal` | `PATCH /v2/pals/{id}` |
107
- | `tavus_conversation_create` | `POST /v2/conversations` |
107
+ | `tavus_conversation_create` | `POST /v2/conversations` with `meeting_url` |
108
108
  | `tavus_conversation_end` | `POST /v2/conversations/{id}/end` |
109
- | `tavus_quickstart` | `POST /v2/pals` + `POST /v2/conversations` |
109
+ | `tavus_quickstart` | `GET /v2/faces` + `POST /v2/pals` |
110
110
  | `tavus_face_list` | `GET /v2/faces` |
111
111
  | `tavus_resource_list` | `GET /v2/{guardrails,objectives,documents,voices,tools}` |
112
112
  | `tavus_scaffold_embed` | none (returns starter HTML/TSX locally) |