fix: correct CLI package names across agent-facing manifests - #455
aisling404 wants to merge 11 commits into
Conversation
AGENTS.md pointed agents at @telnyx/cli to install cli/ (wrong package), and agent.json advertised @telnyx/api-cli, which Telnyx docs deprecate in favor of @telnyx/cli. Fixes: - AGENTS.md: cli/ installs as @telnyx/agent-cli (bin telnyx-agent) - agent.json cli block: @telnyx/api-cli -> @telnyx/cli; command list and global flags re-verified against the installed CLI (v2.0.0), legacy topic matrix replaced with the verified top-level command set - agent.json: new first-class agent_cli block for @telnyx/agent-cli with the 'telnyx-agent capabilities --json' self-description hook - quickstart: examples rewritten with verified syntax (auth login, -d JSON payloads, number order positional arg); leads with the one-command telnyx-agent setup paths - sdks: cli entry repointed to @telnyx/cli; agent_cli entry added Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB
|
@codex review |
![chatgpt-codex-connector[bot]](https://avatars.githubusercontent.com/in/1144995?s=60&v=4){kind=small}
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 36baf75a64
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "video": ["room list", "room get", "room create", "room delete", "session list", "session get", "recording list", "recording get", "recording delete"], | ||
| "voice-profile": ["create", "delete", "get", "list"] | ||
| } | ||
| "package": "@telnyx/cli", |
There was a problem hiding this comment.
After switching the advertised package to @telnyx/cli, several canonical capability examples still use commands from the removed legacy topic matrix: telnyx ai chat (line 62), telnyx storage bucket (line 118), and telnyx 10dlc wizard (line 155). None of those top-level commands appears in the newly verified command list at lines 215-223, so runtime agents that install the package specified here and execute these capability commands will fail; migrate every capability example to the new syntax or point it to telnyx-agent.
AGENTS.md reference: AGENTS.md:L85-L87
Useful? React with 👍 / 👎.
| "install": "npm install -g @telnyx/api-cli", | ||
| "auth": "telnyx auth setup --api-key YOUR_KEY", | ||
| "install": "npm install -g @telnyx/agent-cli @telnyx/cli", | ||
| "auth": "export TELNYX_API_KEY=YOUR_KEY # or: telnyx auth login", |
There was a problem hiding this comment.
If a user follows the advertised telnyx auth login alternative instead of exporting TELNYX_API_KEY, the first quickstart command, telnyx-agent setup-sms, immediately fails because cli/src/commands/setup-sms.ts:58-59 reads only process.env.TELNYX_API_KEY and explicitly rejects a missing variable. Remove this alternative from the shared quickstart or make the setup command consume credentials written by the new CLI.
Useful? React with 👍 / 👎.
Codex review follow-ups: scenario capability examples still used removed
legacy commands (telnyx ai chat, telnyx storage bucket, telnyx 10dlc
wizard) and unverified flags. All examples now verified against the
installed CLIs: assistant create uses -d JSON, ai chat moves to
telnyx-agent ai-chat --message, fax send uses --media, storage uses
bucket create -d, 10dlc points at telnyx-agent setup-10dlc. The
quickstart auth note now states telnyx-agent reads only TELNYX_API_KEY
('telnyx auth login' applies to telnyx commands only).
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB
|
Addressed both Codex findings in 1752ff7:
|
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 1752ff7f3a
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "description": "Register brands and campaigns for US A2P messaging compliance.", | ||
| "guide": "/guides/10dlc-registration.md", | ||
| "cli": "telnyx 10dlc wizard", | ||
| "cli": "telnyx-agent setup-10dlc", |
There was a problem hiding this comment.
Fresh evidence after the earlier legacy-syntax report is that this replacement still cannot execute: setup10dlcCommand rejects invocations without both --phone and --email at cli/src/commands/setup-10dlc.ts:237-244, while the new capability example supplies neither. Any runtime agent invoking this advertised command exits before registering a brand, so include placeholder contact flags or direct users to a complete example.
AGENTS.md reference: AGENTS.md:L85-L87
Useful? React with 👍 / 👎.
| "bin": "telnyx-agent", | ||
| "install": "npm install -g @telnyx/agent-cli", | ||
| "repo": "https://github.com/team-telnyx/ai/tree/main/cli", | ||
| "description": "Agent-friendly CLI: composite setup commands that take you from zero to a working capability in one call. Idempotent by default, --json everywhere.", |
There was a problem hiding this comment.
When an agent retries the highlighted provisioning workflows, this blanket guarantee can create duplicate billable resources: cli/README.md:93-97 explicitly marks setup-ai and setup-10dlc non-idempotent, and setup-porting also creates a new order. Since these commands are advertised immediately below this description, qualify which workflows reuse resources rather than telling runtime agents the CLI is idempotent by default.
AGENTS.md reference: AGENTS.md:L113-L123
Useful? React with 👍 / 👎.
| "telnyx-agent setup-ai --json", | ||
| "telnyx-agent setup-porting --phone-numbers +13125550001 --customer-name 'Acme Corp'" | ||
| ], | ||
| "note": "Run `telnyx-agent capabilities --json` for the authoritative, machine-readable command list." |
There was a problem hiding this comment.
When a runtime agent relies on this note for command discovery, telnyx-agent capabilities --json returns the curated COMPOSITE_COMMANDS array rather than the complete router inventory; the CLI's own cli/README.md:83 explicitly says this catalog is not complete, and executable commands such as fund-account and whatsapp-templates are omitted. Either make the generated list exhaustive or describe it as a selected capability catalog so agents do not conclude that omitted commands are unavailable.
AGENTS.md reference: AGENTS.md:L113-L123
Useful? React with 👍 / 👎.
…lc contact flags Codex round 2: the blanket 'idempotent by default' claim was wrong — cli/README marks setup-ai and setup-10dlc non-idempotent and setup-porting creates a new order each run; the agent_cli description now names which setups reuse resources and which create fresh ones. The capabilities note no longer calls the catalog authoritative (README: 'not the complete router inventory'). The setup-10dlc example now supplies the required --phone/--email flags (setup-10dlc.ts:242 rejects invocations without both). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB
|
Round-2 Codex findings addressed in 3c33d24 — all three were valid:
The same overstated idempotency/authoritative-catalog wording had been copied into the companion PRs; both are corrected too (dotcom-monorepo#3233 @ 1f38e480, developer-docs-mintlify#1266 @ 7256c40). |
|
@codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 3c33d24aba
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "name": "AI Inference", | ||
| "description": "LLM chat completions and embeddings via OpenAI-compatible API.", | ||
| "cli": "telnyx ai chat 'Hello'", | ||
| "cli": "telnyx-agent ai-chat --message 'Hello'", |
There was a problem hiding this comment.
The fresh replacement still fails when a runtime agent invokes the advertised AI-inference example: ai-chat expects each --message value to be a JSON object (as documented by the command example at cli/src/index.ts:1008 and its validation text at cli/src/commands/ai-chat.ts:91-104), while 'Hello' is forwarded to the generated CLI as a scalar and rejected during request parsing. Wrap the prompt as '{"role":"user","content":"Hello"}' so this canonical capability command executes.
AGENTS.md reference: AGENTS.md:L85-L87
Useful? React with 👍 / 👎.
Codex round 3: ai-chat's --message flag requires a JSON message object
(expandMessages rejects plain strings with 'requires a JSON object');
the example now uses '{"role":"user","content":"Hello"}'.
Co-Authored-By: Claude Fable 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 0cc5a04a95
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "description": "Register brands and campaigns for US A2P messaging compliance.", | ||
| "guide": "/guides/10dlc-registration.md", | ||
| "cli": "telnyx 10dlc wizard", | ||
| "cli": "telnyx-agent setup-10dlc --phone +15551234567 --email compliance@example.com", |
There was a problem hiding this comment.
When an agent runs this example as written, setup-10dlc defaults to web opt-in but, without --website, inserts the literal [your opt-in URL] into message_flow and still creates the brand and submits the campaign (cli/src/commands/setup-10dlc.ts:253,278-285,321-361). The command's own README explicitly requires --website for web opt-in so the submitted flow does not contain this placeholder (cli/README.md:423-424); add a representative URL placeholder or select a non-web opt-in method so the canonical command does not create a non-idempotent campaign with invalid compliance evidence.
AGENTS.md reference: AGENTS.md:L85-L87
Useful? React with 👍 / 👎.
Codex round 4: setup-10dlc defaults to web opt-in and, without --website, builds a message_flow containing a literal placeholder URL yet still submits the campaign. The example now includes --website. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB
|
Round-4 finding addressed in e6cd00c — valid: @codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: e6cd00c515
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "version": "2.0.0", | ||
| "bin": "telnyx", | ||
| "install": "npm install -g @telnyx/cli", | ||
| "repo": "https://github.com/team-telnyx/telnyx-cli", |
There was a problem hiding this comment.
When a runtime agent follows this newly added repository link for @telnyx/cli v2, it lands on the incompatible Go CLI documented in skills/telnyx-cli/SKILL.md:21-65, whose plural resource syntax (telnyx messages create) differs from the singular npm commands advertised here; cli/src/platform-release.ts:3-24 likewise identifies that repository's releases as Go CLI v0.27. Point this field to the npm CLI's actual source/docs, or advertise the Go CLI and its matching install and commands instead.
AGENTS.md reference: AGENTS.md:L85-L87
Useful? React with 👍 / 👎.
Codex round 5: github.com/team-telnyx/telnyx-cli hosts the Go platform CLI (plural-resource syntax; the agent CLI vendors its releases in platform-release.ts), which is not the npm @telnyx/cli package this block documents. Link the npm page instead and note the distinction. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB
|
Round-5 finding addressed in 8dd4717 — valid: github.com/team-telnyx/telnyx-cli hosts the Go platform CLI (plural-resource syntax; Reviewer heads-up beyond this PR's scope: the npm package's own @codex review |
|
Codex Review: Didn't find any major issues. Nice work! Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
Real E2E validation against a live Telnyx account showed the npm @telnyx/cli is human-oriented and unfit for the agent manifest's examples: auth requires interactive 'telnyx auth login' (TELNYX_API_KEY is not read), number search and call dial drop into interactive prompts, message send loses the from field, and the generic CRUD layer (messaging-profile/assistant/bucket create) fails with 'Resource not found'. Every scenario and quickstart example is now a telnyx-agent command validated end-to-end (setup-sms bought+wired a real number and proved reuse on re-run; send-sms queued a real SMS; ai-chat returned a completion; create-ai-assistant works with Qwen/Qwen3-235B-A22B — the Llama model id is rejected by the assistants API). The storage scenario loses its CLI example (the agent CLI has no bucket op). The @telnyx/cli block shrinks to a minimal pointer with an honest interactivity note. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB
Live E2E validation (real account, real money)Installed both published packages and executed the manifest's commands against a live Telnyx account. Result: every runnable example is now a telnyx-agent (@telnyx/agent-cli 0.4.2) — validated live ✅
🐛 Bug found in @telnyx/agent-cli 0.4.2 (separate issue-worthy)
npm @telnyx/cli 2.0.0 — why its examples were removed ❌
Resources created during validation (owner may want to release): +16073051147, +16362029859, +19252588249 (orphaned by the setup-ai bug), 2 messaging profiles, 1 Call Control App, 1 Verify profile, 2 AI assistants. @codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: cdb1044408
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "name": "Zero to AI assistant on a phone number (one command)", | ||
| "commands": [ | ||
| "telnyx ai chat 'What can Telnyx do?'" | ||
| "telnyx-agent setup-ai --json" |
There was a problem hiding this comment.
With the advertised @telnyx/agent-cli 0.4.2, this command creates an assistant and purchases a number before attempting to wire them using phone-numbers update --force; the vendored Go CLI does not support that flag (cli/src/commands/setup-ai.ts:99-104), so the quickstart exits unsuccessfully after creating billable orphaned resources. Remove setup-ai from the quickstart and highlights until its wiring step is fixed, or replace it with the validated create-ai-assistant example.
AGENTS.md reference: AGENTS.md:L115-L123
Useful? React with 👍 / 👎.
Codex round 7, matching the live E2E result: @telnyx/agent-cli 0.4.2's setup-ai creates an assistant and buys a number, then fails wiring (--force rejected by the vendored Go CLI), leaving billable orphans. Remove it from the quickstart and highlights, keep the validated create-ai-assistant example, and note the v0.4.2 breakage in the agent_cli description. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB
|
Round-7 finding addressed in b4afdd9 — valid (and consistent with the live E2E result above): @codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: b4afdd97bf
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "name": "@telnyx/cli", | ||
| "install": "npm install -g @telnyx/cli", | ||
| "docs": "https://github.com/team-telnyx/telnyx-cli" |
There was a problem hiding this comment.
When a runtime agent follows the sdks.cli.docs link, this entry now names and installs @telnyx/cli but still directs the agent to the incompatible Go platform CLI repository. Fresh evidence after the prior fix is that cli.docs was corrected to the npm page while this separate SDK entry was left unchanged; the linked CLI uses the Go installation and plural-resource syntax documented in skills/telnyx-cli/SKILL.md:21-65. Point this field to the same npm documentation as the top-level CLI entry.
AGENTS.md reference: AGENTS.md:L113-L123
Useful? React with 👍 / 👎.
Codex round 8: same mismatch as the cli block had — the sdks entry names npm @telnyx/cli but linked the incompatible Go platform CLI repo. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB
|
Round-8 finding addressed in 23f46bd — valid, same npm-vs-Go mismatch as round 5 but in the @codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 23f46bd7a2
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "highlights": [ | ||
| "telnyx-agent status", | ||
| "telnyx-agent setup-sms --json", | ||
| "telnyx-agent setup-voice --json", |
There was a problem hiding this comment.
When a runtime agent follows this new highlight on an account with an outbound voice profile but no reusable app, setup-voice silently substitutes https://example.com/webhook (cli/src/commands/setup-voice.ts:48-49), provisions a number, and reports ready: true even though call events and inbound-call handling go to the placeholder rather than the user's application. Include a representative --webhook-url placeholder in the highlighted command or stop presenting this invocation as a working setup.
AGENTS.md reference: AGENTS.md:L113-L123
Useful? React with 👍 / 👎.
Codex round 9: without --webhook-url, setup-voice silently wires the Call Control App to https://example.com/webhook and reports ready:true while call events go to the placeholder (live E2E also showed such an app is rejected for dialing). The highlight now shows the flag. Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB
|
Round-9 finding addressed in 11a0d3c — valid, and it matches the live E2E: @codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 11a0d3c860
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "name": "Zero to sending SMS (one command)", | ||
| "commands": [ | ||
| "telnyx messaging-profile create --name 'My Profile' --json", | ||
| "telnyx number search --country US --type long_code --json", | ||
| "telnyx number order --phone-number +1XXXXXXXXXX --json", | ||
| "telnyx message send --from +1XXXXXXXXXX --to +15559876543 --text 'Hello from Telnyx!'" | ||
| "telnyx-agent setup-sms --json" |
There was a problem hiding this comment.
For the default US setup, this command only creates a messaging profile, buys a local number, and assigns it; it never performs the 10DLC registration that guides/sms-messaging.md:5-10 requires for US A2P traffic, even though setup-sms reports ready: true. An agent following this advertised “Zero to sending SMS” path can therefore purchase a number but still have its messages rejected as not_10dlc_registered; add the registration/assignment workflow or describe this as provisioning only.
AGENTS.md reference: AGENTS.md:L115-L123
Useful? React with 👍 / 👎.
Codex round 10: setup-sms provisions a profile and number and reports ready:true, but US A2P delivery also requires 10DLC registration — without it messages can be rejected as not_10dlc_registered. The quickstart block is renamed to 'Provision for SMS' and both SMS blocks carry a note pointing at setup-10dlc (and its carrier fees). Co-Authored-By: Claude Fable 5 <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB
|
Round-10 finding addressed in f5bc1b7 — valid: @codex review |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: f5bc1b71b5
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
| "telnyx-agent buy-phone-number --phone-number +1XXXXXXXXXX --messaging-profile-id PROFILE_ID --json", | ||
| "telnyx-agent send-sms --from +1XXXXXXXXXX --to +15559876543 --text 'Hello from Telnyx!' --json" | ||
| ], | ||
| "note": "For US A2P traffic, register 10DLC before sending (telnyx-agent setup-10dlc); see the 10dlc capability above." |
There was a problem hiding this comment.
Fresh evidence in this new note is that the referenced capability example omits --phone-number-id; setup-10dlc only executes its phone-number assignment step when that flag is supplied (cli/src/commands/setup-10dlc.ts:258,372-389), while the 10DLC guide identifies assignment as a required third step. An agent following either SMS quickstart can therefore create a billable brand and campaign yet leave the purchased sender outside that campaign, so the subsequent US A2P send can still be rejected. Include the purchased E.164 number in the registration invocation and direct the agent to wait for campaign approval.
AGENTS.md reference: AGENTS.md:L85-L87
Useful? React with 👍 / 👎.
1 participant
Summary
An agent following this repo's own manifests can currently install the wrong CLI:
AGENTS.mdtold agents to installcli/as@telnyx/cli— that's the general CLI (telnyx-cli repo), not the agent CLI. The right package is@telnyx/agent-cli(bintelnyx-agent).agent.jsonadvertised@telnyx/api-cli, which Telnyx's own docs deprecate in favor of@telnyx/cli— and its command matrix + quickstart used the legacy CLI's syntax, which does not work on the current one (e.g.auth setup,--nameflags,telnyx ai chat,telnyx storage bucket).Changes
cli/row now installs@telnyx/agent-cli(bintelnyx-agent).cliblock: repointed to@telnyx/cliv2.0.0; global flags and the full top-level command list re-verified against the installed CLI (npx @telnyx/cli --helpand per-command--help). Legacy topic matrix removed; a note points attelnyx resources/--helpfor drill-down so the manifest can't drift as easily.agent_cliblock (new): first-class entry for@telnyx/agent-cliwith install, bin, repo, and thetelnyx-agent capabilities --jsonself-description hook.auth login,-d '{...}'payloads, positionalnumber order); now leads with the one-commandtelnyx-agent setup-sms/setup-aipaths.clientry fixed;agent_clientry added.Verification
Every command in the new
cliblock and quickstart was checked against the actual installed@telnyx/cli2.0.0 (--helpoutput per subcommand).agent.jsonpassespython3 -m json.tool. The assistant example's model id matches the one used 5× in this repo's guides (meta-llama/Llama-3.3-70B-Instruct).Not included
npm deprecate @telnyx/api-cli— needs frontend-squad npm rights; recommended follow-up.🤖 Generated with Claude Code
https://claude.ai/code/session_01LFvhpJqZ1F3szEBV3WZnYB