Skip to main content

valet

Use when the user wants to manage Valet agents, channels, connectors, organizations, or environment variables (secrets and plain config) via the valet CLI. Handles creation, deployment, linking, teardown, and all multi-step workflows. Also use when asked to "create an agent", "deploy an agent", "...

Install this skill

or
0/100

Security score

The valet skill was audited on Aug 18, 2026 and we found 54 security issues across 7 threat categories, including 1 critical. Review the findings below before installing.

Categories Tested

Security Issues

high line 30

Piping content to sh shell

SourceSKILL.md
28
29```sh
30curl -fsSL https://valet.dev/install.sh | sh
31```
32
medium line 30

Curl to non-GitHub URL

SourceSKILL.md
28
29```sh
30curl -fsSL https://valet.dev/install.sh | sh
31```
32
medium line 674

Curl to non-GitHub URL

SourceSKILL.md
672
673# Pass a secret as a CLI argument using {{}} syntax
674valet exec -a my-agent API_KEY -- curl -H "Authorization: Bearer {{API_KEY}}" https://api.example.com
675
676# Multiple secrets in one command
medium line 704

Curl to non-GitHub URL

SourceSKILL.md
702```
703# WRONG — $API_KEY is not set in your shell
704curl https://api.example.com/data?key=$API_KEY
705
706# CORRECT — valet exec injects the secret (explicit secrets mode)
medium line 707

Curl to non-GitHub URL

SourceSKILL.md
705
706# CORRECT — valet exec injects the secret (explicit secrets mode)
707valet exec -a my-agent API_KEY -- curl https://api.example.com/data?key={{API_KEY}}
708
709# OR use connector mode if you have a command connector configured
medium line 748

Curl to non-GitHub URL

SourceSKILL.md
746# Test the endpoint is reachable and the token works:
747valet exec -a my-agent API_TOKEN -- \
748 curl -s -o /dev/null -w "%{http_code}" -H "Authorization: {{API_TOKEN}}" https://mcp.example.com/mcp
749```
750
medium line 1079

Access to home directory dotfiles

SourceSKILL.md
1077
10781. Convert the current working directory to the Claude projects path:
1079 `~/.claude/projects/-<cwd-with-slashes-replaced-by-dashes>/`
1080 Example: `/Users/me/Developer/my-project` → `~/.claude/projects/-Users-me-Developer-my-project/`
10812. Find the active session log:
medium line 1080

Access to home directory dotfiles

SourceSKILL.md
10781. Convert the current working directory to the Claude projects path:
1079 `~/.claude/projects/-<cwd-with-slashes-replaced-by-dashes>/`
1080 Example: `/Users/me/Developer/my-project` → `~/.claude/projects/-Users-me-Developer-my-project/`
10812. Find the active session log:
1082 ```bash
low line 1083

Access to home directory dotfiles

SourceSKILL.md
10812. Find the active session log:
1082 ```bash
1083 ls -t ~/.claude/projects/-<path>/*.jsonl | head -1
1084 ```
1085
medium line 1099

Access to home directory dotfiles

SourceSKILL.md
1097For large sessions (>20 user prompts): sample the first 3 and last 3 user prompts to keep context manageable.
1098
1099Also check `~/.claude/projects/<project-path>/sessions-index.json` for `summary` and `firstPrompt` fields matching the session ID (derived from the JSONL filename).
1100
1101If the session is empty (no user prompts besides the learn trigger), inform the user and stop.
medium line 3

Webhook reference

SourceSKILL.md
1---
2name: valet
3description: Use when the user wants to manage Valet agents, channels, connectors, organizations, or environment variables (secrets and plain config) via the valet CLI. Handles creation, deployment, linking, teardown, and all multi-step workflows. Also use when asked to "create an agent", "deploy an agent", "design an agent", "build me an agent that...", "create a connector", "set up a webhook", or anything involving the Valet platform or any request to create and deploy AI agents. Also use when asked to "learn from this session", "capture this workflow", "save this as an agent", "make this repeatable", or when writing SOUL.md files.
4---
5
medium line 116

Webhook reference

SourceSKILL.md
114- **Organization**: A team workspace that owns agents, connectors, channels, and env vars. All agents belong to an org — the default org is used when `--org` is omitted.
115- **Connector**: An MCP server or CLI tool that provides capabilities to agents. Types: `mcp-server` (MCP tools via client) and `command` (CLI with secret injection). Transports: `stdio`, `sse`, `streamable-http`.
116- **Channel**: A message entry point for agents. Types: `webhook`, `slack`, `telegram`, `heartbeat`, `cron`, `console`, `mcp`. Each channel has a session strategy and a prompt path.
117- **Env var**: A named value scoped to an org or agent, encrypted at rest, with one of two kinds. **Secret** (the default) is a credential the agent can use through connectors and channels but never sees. **Plain** is agent-visible configuration delivered to the agent's environment, readable as `$NAME`. Either kind is referenced with `{{NAME}}` template syntax in connector and channel configurations. Agent-scoped env vars override org-scoped env vars of the same name.
118- **Catalog**: A Valet-curated library of well-known connector and channel definitions. Browse with `valet connectors catalog` or `valet channels catalog`. Add from the catalog instead of configuring from scratch.
medium line 141

Webhook reference

SourceSKILL.md
1391. **Catalog first**: Check `valet connectors catalog` or `valet channels catalog` before creating from scratch. Catalog entries handle transport, commands, and secret slots automatically.
1402. **Reuse existing**: Check `valet connectors --org <org>` or `valet channels --org <org>` for resources that already provide what you need. Attach rather than duplicate.
1413. **Org-scoped by default**: Always create connectors, channels, and env vars at the org level (`--org`). Org-scoped resources can be attached to any agent in the org, so a single `GITHUB_TOKEN` secret, `github` MCP connector, or `github-webhook` channel is reusable across every agent — no duplication, one place to rotate credentials. **Only drop to `--agent` when you have a concrete reason the resource cannot be shared** (e.g., per-agent rate limits, distinct credentials for the same service, a one-off test agent). When in doubt, use `--org` and attach. (Slack is a special case — see the Channels section.)
1424. **Env vars at org level by default**: Setting an env var with `--org` makes it available to every org-scoped connector and channel, and org-scoped plain vars reach every agent in the org. Any agent that later attaches those connectors/channels automatically inherits access — no duplication. Agent-scoped env vars override org-scoped ones of the same name when a specific agent needs a different value.
1435. **Verify before finalizing**: After the agent exists, test every
medium line 371

Webhook reference

SourceSKILL.md
369## Channels
370
371Channels are message entry points for agents. **Default to `--org` when creating webhook channels** — one org-scoped webhook can be attached to multiple agents (each with its own `--events` filter), reusing the same webhook URL and signing secret. Use `--agent` when the channel truly belongs to one agent (cron and heartbeat schedules, Telegram bots).
372
373**Slack is a two-step special case.** The org-level Slack channel (`valet channels create slack --org <org>`) authorizes Valet to create apps in your Slack workspace — one per org, a one-time prerequisite, not a reusable channel. Each agent that needs to appear in Slack then gets its **own** per-agent Slack channel (`valet channels create slack --agent <name>` or `valet channels attach slack --agent <name>`), which provisions a dedicated Slack app with its own bot identity. See "Create a Slack channel" below for the full flow.
medium line 375

Webhook reference

SourceSKILL.md
373**Slack is a two-step special case.** The org-level Slack channel (`valet channels create slack --org <org>`) authorizes Valet to create apps in your Slack workspace — one per org, a one-time prerequisite, not a reusable channel. Each agent that needs to appear in Slack then gets its **own** per-agent Slack channel (`valet channels create slack --agent <name>` or `valet channels attach slack --agent <name>`), which provisions a dedicated Slack app with its own bot identity. See "Create a Slack channel" below for the full flow.
374
375Follow the Resource Creation Principles above — the catalog encodes signing schemes and service-specific behaviors for webhook channels.
376
377**Always pass an explicit name when creating a channel that has a user-visible identity** (Slack `--bot-name`, Telegram bot name, etc.). Auto-defaults from the agent name are not reliable across orgs and surfaces — supply the flag yourself rather than letting it resolve server-side. For Slack specifically, this means passing `--bot-name <display-name>` on every `valet channels create slack --agent ...` and `valet channels attach slack --agent ...` call.
low line 399

Webhook reference

SourceSKILL.md
397```
398valet env set GITHUB_WEBHOOK_SECRET=whsec_abc123 --org acme
399valet channels create github-webhook --org acme
400```
401
low line 419

Webhook reference

SourceSKILL.md
417Example:
418```
419valet channels attach github-webhook --agent my-reviewer --events pull_request,issue_comment
420valet channels attach slack --agent my-agent --bot-name my-bot
421valet channels detach slack --agent my-agent
medium line 424

Webhook reference

SourceSKILL.md
422```
423
424### Create a webhook channel
425
426```
low line 427

Webhook reference

SourceSKILL.md
425
426```
427valet channels create webhook [name] \
428 [--agent <agent-name>] [--org <org>] \
429 [--verify <scheme>]
medium line 432

Webhook reference

SourceSKILL.md
430```
431
432Verification schemes: `hmac-sha256` (default), `slack`, `stripe`, `svix`, `static-token`, `none`. Key flags: `--secret-name` (name of an existing secret from `valet env` to use instead of auto-generating; required for `slack`, `stripe`, and `svix`), `--signature-header` (not used with `slack` or `svix`), `--delivery-key-header`, `--delivery-key-path`, `--prompt`. For `hmac-sha256` and `static-token`, a managed secret is auto-generated if `--secret-name` is omitted. The `slack` scheme implements Slack's Events API signing protocol and handles `url_verification` challenges automatically. Run `valet channels create webhook --help` for full details.
433
434The command prints the **webhook URL**, **signing secret**, and, when
medium line 434

Webhook reference

SourceSKILL.md
432Verification schemes: `hmac-sha256` (default), `slack`, `stripe`, `svix`, `static-token`, `none`. Key flags: `--secret-name` (name of an existing secret from `valet env` to use instead of auto-generating; required for `slack`, `stripe`, and `svix`), `--signature-header` (not used with `slack` or `svix`), `--delivery-key-header`, `--delivery-key-path`, `--prompt`. For `hmac-sha256` and `static-token`, a managed secret is auto-generated if `--secret-name` is omitted. The `slack` scheme implements Slack's Events API signing protocol and handles `url_verification` challenges automatically. Run `valet channels create webhook --help` for full details.
433
434The command prints the **webhook URL**, **signing secret**, and, when
435applicable, **managed secret name**. Report the URL and managed secret name.
436Do not repeat the raw signing secret in the response. Tell the user to copy
medium line 751

Webhook reference

SourceSKILL.md
749```
750
751Do not test a webhook by sending an arbitrary request to a production
752endpoint. Use the provider's documented test-event flow or a dedicated test
753endpoint. If either action can trigger real work, get the user's approval
low line 794

Webhook reference

SourceSKILL.md
7924. Add channels from the catalog at the org level (for webhooks):
793 ```
794 valet channels create github-webhook --org acme
795 ```
796
low line 806

Webhook reference

SourceSKILL.md
804 valet agents create my-agent --org acme \
805 --attach-connector github \
806 --attach-channel github-webhook
807 ```
808 Or attach after creation:
low line 811

Webhook reference

SourceSKILL.md
809 ```
810 valet connectors attach github --agent my-agent
811 valet channels attach github-webhook --agent my-agent --events pull_request
812 ```
813
low line 876

Webhook reference

SourceSKILL.md
874# Detach org resources (they remain available for other agents)
875valet connectors detach github --agent my-agent
876valet channels detach github-webhook --agent my-agent
877
878# Destroy agent-scoped resources
medium line 932

Webhook reference

SourceSKILL.md
930
931Ask about the trigger if not already clear:
932- Webhook — event-driven (email, push, form submission)
933- Prompt — user sends a message via `valet run` or console
934
medium line 1009

Webhook reference

SourceSKILL.md
1007 - [ ] Guardrails has both Always and Never subsections
1008 - [ ] No hardcoded IDs that should be `<placeholder>`s
1009 - [ ] Channel files have Scope section if webhook-driven
1010 - [ ] Channel files include webhook payload location instruction
1011 - [ ] No secrets or API keys in any file
medium line 1010

Webhook reference

SourceSKILL.md
1008 - [ ] No hardcoded IDs that should be `<placeholder>`s
1009 - [ ] Channel files have Scope section if webhook-driven
1010 - [ ] Channel files include webhook payload location instruction
1011 - [ ] No secrets or API keys in any file
1012 - [ ] AGENTS.md documents all setup without containing secret values
medium line 1034

Webhook reference

SourceSKILL.md
1032 --env KEY={{SECRET_NAME}}
1033 ```
103410. Set up channels — **check the catalog first** for webhook channels:
1035 ```
1036 valet channels catalog
low line 1041

Webhook reference

SourceSKILL.md
1039 Or create directly:
1040 ```
1041 valet channels create webhook <channel-name> --agent <agent-name>
1042 ```
104311. Create and deploy the agent, attaching org resources:
medium line 1119

Webhook reference

SourceSKILL.md
1117Ask clarifying questions (skip any with obvious answers from the session):
1118
11191. **Trigger**: What should invoke this agent? Propose a draft based on the first user prompt — webhook or prompt?
11202. **Scope**: Does the extracted objective + tool list capture the full scope, or should it be narrowed/expanded?
11213. **Corrections**: Surface each detected correction and ask whether the agent should always follow the corrected approach.
medium line 1190

Webhook reference

SourceSKILL.md
1188### Optional sections
1189
1190Add as needed: **Target Channel**, **Environment Requirements**, **Webhook Scope Rule**, **Skills Used**, **MEMORY.md Format**.
1191
1192### Synthesis rules
medium line 1214

Webhook reference

SourceSKILL.md
1212- Missing Workflow — Purpose without steps leaves the agent guessing
1213- Hardcoded values that should be `<placeholder>`s
1214- No scope boundary for webhook agents (see Writing Channel Files)
1215- Using the wrong command name for command connectors — the connector must be named after the CLI command (e.g., `agentmail` not `agentmail-cli`), and SOUL.md must reference that same name
1216
medium line 1221

Webhook reference

SourceSKILL.md
1219Channel files tell the agent what to do when a message arrives. They are instructions TO the agent, written as direct imperatives.
1220
1221### Webhook payload location (critical)
1222
1223The JSON webhook payload is appended inline after the channel file in the user message. Every channel file **must** start with:
medium line 1223

Webhook reference

SourceSKILL.md
1221### Webhook payload location (critical)
1222
1223The JSON webhook payload is appended inline after the channel file in the user message. Every channel file **must** start with:
1224
1225```
low line 1226

Webhook reference

SourceSKILL.md
1224
1225```
1226The JSON webhook payload is appended directly after these instructions
1227in the user message. Parse it inline — do not fetch, list, or search
1228for the payload elsewhere. Do NOT use tools to read the payload.
low line 1246

Webhook reference

SourceSKILL.md
1244# New Email Received
1245
1246The JSON webhook payload is appended directly after these instructions
1247in the user message. Parse it inline — do not fetch, list, or search
1248for the payload elsewhere. Do NOT use tools to read the payload.
low line 1250

Webhook reference

SourceSKILL.md
1248for the payload elsewhere. Do NOT use tools to read the payload.
1249
1250You received a webhook for a single new email.
1251
1252## Scope
medium line 1265

Webhook reference

SourceSKILL.md
1263### Reinforcing scope in SOUL.md
1264
1265For webhook-driven agents, add to SOUL.md:
1266
1267```markdown
low line 1268

Webhook reference

SourceSKILL.md
1266
1267```markdown
1268## Webhook Scope Rule
1269
1270When you receive a webhook, your scope of work is defined by the
low line 1270

Webhook reference

SourceSKILL.md
1268## Webhook Scope Rule
1269
1270When you receive a webhook, your scope of work is defined by the
1271identifiers in the payload. Use any tools to fully understand and act
1272on that specific content, but do not act on unrelated content.
medium line 1398

Webhook reference

SourceSKILL.md
1396- Write in plain English — describe requirements as nouns with reasons, not CLI commands
1397- Be specific about secrets — include required scopes/permissions and where to obtain them
1398- Include external setup steps (OAuth apps, cloud consoles, webhook registrations, etc.)
1399- Omit sections that don't apply. Write this file last.
1400
low line 1432

Webhook reference

SourceSKILL.md
1430 - role: trigger
1431 title: "Your team asks a question in Slack."
1432 body: "Someone posts 'where do we handle the Stripe webhook retry logic?'"
1433 catalog: slack-webhook
1434 - role: action
low line 1433

Webhook reference

SourceSKILL.md
1431 title: "Your team asks a question in Slack."
1432 body: "Someone posts 'where do we handle the Stripe webhook retry logic?'"
1433 catalog: slack-webhook
1434 - role: action
1435 title: "AskADev reads your code."
medium line 1564

Webhook reference

SourceSKILL.md
1562| hero | *"AskADev is an AI-powered Slack assistant that answers code questions."* | *"A Slack bot that reads your code before it answers."* | Bad leads with category + buzzword. Good leads with behavior and names the hook ("before it answers"). |
1563| subheadline | *"Uses GitHub MCP and Slack MCP to provide intelligent responses to developer questions."* | *"Ask a question about a GitHub repo in Slack. AskADev researches the actual code and commit history, then replies in-thread."* | Bad names the plumbing. Good names the action and the reward. |
1564| trigger title | *"Webhook event received"* | *"A PR is opened."* | Bad names the mechanism. Good names what happened in the user's world. |
1565| action title | *"Diff analysis"* | *"Code Reviewer reads the diff."* | Bad is a noun phrase. Good is a sentence with a subject and verb. |
1566| outcome title | *"Review submitted"* | *"Inline comments — or an approve."* | Bad is passive mechanism. Good names the two concrete outputs. |
medium line 1567

Webhook reference

SourceSKILL.md
1565| action title | *"Diff analysis"* | *"Code Reviewer reads the diff."* | Bad is a noun phrase. Good is a sentence with a subject and verb. |
1566| outcome title | *"Review submitted"* | *"Inline comments — or an approve."* | Bad is passive mechanism. Good names the two concrete outputs. |
1567| step body | *"The agent processes the incoming webhook payload and performs configured actions."* | *"Checks correctness, security, maintainability, and test coverage. Reads full files when context matters."* | Bad is generic. Good enumerates specifics. |
1568| ui blurb | *"Connect your GitHub account to give the agent access to your repositories."* | *"AskADev reads your code when it answers — like a new hire would."* | Bad could belong to any GitHub connector. Good is about *this* agent's use of GitHub. |
1569| ui done_note | *"Successfully connected"* | *"Listening in #engineering"* | Bad is a status. Good is the concrete outcome the user wanted. |
low line 1586

Webhook reference

SourceSKILL.md
1584
1585channels:
1586 - catalog: slack-webhook
1587 description: "Receives Slack messages and app mentions"
1588 ui:
low line 1664

Webhook reference

SourceSKILL.md
1662 SOUL.md # Agent identity and behavior (required)
1663 valet.yaml # From valet new; required for catalog publication
1664 channels/ # Channel files (for webhook/trigger-driven agents)
1665 <channel-name>.md
1666 skills/ # Agent-scoped skill documentation (optional)
medium line 1700

Webhook reference

SourceSKILL.md
1698- When authoring a `valet.yaml` for a catalog-published agent, follow "Authoring the agent story". Run `valet manifests validate` after every edit — length caps and the 3-step role order are non-negotiable.
1699- For destructive commands (`destroy`, `remove`, `revoke`), always confirm with the user first.
1700- When creating webhook channels, report the webhook URL and managed secret
1701 name. Do not repeat the raw signing secret. When writing channel files,
1702 include the payload location instruction.
high line 185

Crontab modification

SourceSKILL.md
183```
184
185Use `type` (mutually exclusive with `catalog`) to declare inline channels. Supported fields: `schedule` (human-readable), `cron` (raw crontab expression), `every` (heartbeat interval), `timezone` (IANA timezone, default UTC). The `mcp` type takes none of these — it stands the agent up as an MCP server, materialized at deploy; mint a bearer token afterward with `valet channels token mint <channel> <name>` to connect a client. Run `valet agents create --help` for all options.
186
187### Link a directory
medium line 484

Crontab modification

SourceSKILL.md
482```
483valet channels create cron [name] --agent <agent-name> --schedule "every day at 9am"
484# Or with a raw crontab expression:
485valet channels create cron [name] --agent <agent-name> --cron "0 9 * * *"
486```
critical line 1678

Silent execution instruction

SourceSKILL.md
1676
1677- Run commands with the terminal tool appropriate to the operating system.
1678- **Be explanatory**: Before running any valet command, briefly tell the user *what* you're about to do and *why*. Don't silently execute commands — the user should always understand the purpose of each step.
1679- **Installation guardrails**: Follow the Installation section strictly.
1680 Use the official direct installer for new installations and `valet
high line 30

Curl pipe to interpreter

SourceSKILL.md
28
29```sh
30curl -fsSL https://valet.dev/install.sh | sh
31```
32
Scanned on Aug 18, 2026
View Security Dashboard
Installation guide →
Rate this skill
Categorydevelopment
UpdatedSeptember 20, 2026
claudeclaude-codefrontenddesignplaywrightstripedocxgitapitestingdevopsbackend
valetdotdev/skills