@usecoil/skill-claude 0.1.0 → 0.1.2

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
package/README.md CHANGED
@@ -1,11 +1,13 @@
1
1
  # Coil Claude Code Skill
2
2
 
3
- Installable Claude Code skill package for Coil agent workflows. It verifies the
4
- installed CLI and installs the compatible public `@usecoil/cli` package when
3
+ Installable Claude Code skill package for Coil agent workflows. The command
4
+ below is the exact install form for this immutable package. Confirm registry
5
+ availability against the reviewed release manifest before use. It verifies the
6
+ installed CLI and installs the compatible `@usecoil/cli` package when
5
7
  `coil` is not already available:
6
8
 
7
9
  ```bash
8
- npx --yes @usecoil/skill-claude@0.1.0 --base-url https://www.usecoil.com
10
+ npx --yes @usecoil/skill-claude@0.1.2 --base-url https://www.usecoil.com
9
11
  coil config set-base-url https://www.usecoil.com --profile prod
10
12
  printf '%s' "$COIL_API_KEY" | coil auth login --profile prod --key -
11
13
  coil --profile prod agent-context --json
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@usecoil/skill-claude",
3
- "version": "0.1.0",
3
+ "version": "0.1.2",
4
4
  "description": "Claude Code skill package for Coil agent workflows",
5
5
  "type": "module",
6
6
  "bin": {
@@ -28,7 +28,7 @@
28
28
  },
29
29
  "coilCompatibility": {
30
30
  "cli": "0.1.x",
31
- "preferredVersion": "0.1.0"
31
+ "preferredVersion": "0.1.2"
32
32
  },
33
33
  "coilRuntime": {
34
34
  "name": "claude",
package/skill/SKILL.md CHANGED
@@ -16,7 +16,7 @@ it can do:
16
16
 
17
17
  ```bash
18
18
  if ! command -v coil >/dev/null 2>&1; then
19
- npm install --global @usecoil/cli@0.1.0
19
+ npm install --global @usecoil/cli@0.1.2
20
20
  fi
21
21
  coil --version
22
22
  ```
@@ -77,7 +77,7 @@ The runtime-specific installer or registry handles placement; Coil operations
77
77
  always use the same public JSON CLI:
78
78
 
79
79
  ```bash
80
- npm install --global @usecoil/cli@0.1.0
80
+ npm install --global @usecoil/cli@0.1.2
81
81
  coil config set-base-url https://www.usecoil.com --profile prod
82
82
  printf '%s' "$COIL_API_KEY" | coil auth login --profile prod --key -
83
83
  coil --profile prod agent-context --json
@@ -98,6 +98,32 @@ other runtimes may use their own secret manager. Do not add `COIL_ORG_ID`.
98
98
  - Diagnostics and errors go to stderr.
99
99
  - Global flags: `--profile`, `--base-url`, `--json`, `--output-version`.
100
100
 
101
+ ## Human action handoffs
102
+
103
+ When a raw Coil API success response includes `meta.human_action`, or a blocked
104
+ API error includes `error.details.human_action`, relay the returned action to
105
+ the user. In CLI `--json` success output, object results and V2 list envelopes
106
+ expose the same action as a top-level `human_action` field. The V1 automations
107
+ result list keeps its array shape and adds `human_action` to the matching item.
108
+ CLI errors preserve it under
109
+ `error.details.human_action`. State whether it is required or recommended,
110
+ include the exact `url` verbatim, and use its human-readable `label`. Do not
111
+ say only “go to the dashboard,” fabricate a URL, or direct draft review to the
112
+ public `/recipes` concept pages. If no specific recipe ID is available, use the
113
+ server-approved `/dashboard/recipes/review` queue. Treat opening the URL as
114
+ navigation only; the destination still authenticates and authorizes the human.
115
+
116
+ Older Actor Lab responses may also contain `approval_url`. Preserve and relay
117
+ that server-provided URL when present, while preferring the standardized
118
+ `human_action` object for machine reasoning.
119
+
120
+ Automation handoffs use the server-provided absolute `/automations` URL. After
121
+ an agent creates or validates a draft, relay the required publication action
122
+ verbatim: the human org admin must review and publish the draft. If a run or
123
+ result has `reconciliation_required`, relay the required reconciliation action
124
+ verbatim and do not retry the webhook. A policy-review action means a human
125
+ must decide the governance policy; it is not an instruction to weaken policy.
126
+
101
127
  ## Common Workflows
102
128
 
103
129
  ### Orient
@@ -112,6 +138,24 @@ coil --profile prod auth status --json
112
138
 
113
139
  Use the server-derived activation state before starting a new workspace flow. Provider-backed runs require a healthy saved connection and explicit spend confirmation.
114
140
 
141
+ `coil activation status --json` may include a server-provided `human_action`.
142
+ State whether it is required or recommended and relay its absolute `url`
143
+ verbatim. Do not map `nextAction` to a guessed dashboard route. `start_run`
144
+ still needs explicit provider-spend confirmation; a human-action URL is only
145
+ navigation and does not grant approval or authority.
146
+
147
+ Machine recipe creation, draft updates, and template installation return a required
148
+ `human_action` for the specific `/dashboard/recipes/{recipeId}` detail surface.
149
+ Successful machine validation of a draft returns the same action. Relay its
150
+ absolute `url` verbatim. A machine publish `403` preserves the scoped action in
151
+ `error.details.human_action` when the draft exists; do not treat the URL as
152
+ approval or authority.
153
+
154
+ For `recipes create --json`, preserve every returned recipe field and relay the
155
+ top-level `human_action` when the server provides it. If the response contains
156
+ only `data` or omits valid metadata, report the unchanged recipe result. Never
157
+ infer or fabricate a dashboard URL or human action.
158
+
115
159
  ```bash
116
160
  coil --profile prod activation status --json
117
161
  coil --profile prod marketplace templates --json
@@ -147,7 +191,7 @@ coil --profile prod recipe-runs cancel <run-id> --json
147
191
  coil --profile prod recipe-runs retry <run-id> --json
148
192
  ```
149
193
 
150
- Publication and ambiguous-effect reconciliation require a human organization admin. Retry/cancel behavior follows the durable run state returned by the API.
194
+ Publication and ambiguous-effect reconciliation require a human organization admin. If a run has an unresolved provider effect, relay the server-provided `human_action.url` from the view/wait/watch state (or `error.details.human_action` from a machine reconcile denial) verbatim. The URL opens `/dashboard/recipe-runs/{runId}` for evidence review; it is navigation only and does not prove provider acceptance or authorize a decision. Retry/cancel behavior follows the durable run state and `next_actions`, not the presence of a URL.
151
195
 
152
196
  ### Scrapes
153
197
 
@@ -175,6 +219,14 @@ coil --profile prod leads emails <scrape-id>
175
219
 
176
220
  API-key agents can create draft automations. Human org admins publish drafts after validation.
177
221
 
222
+ Creation and successful draft validation may return a required `human_action`
223
+ with an absolute `/automations` URL. Tell the user that the draft is ready but
224
+ requires human admin review and publication, and relay `human_action.url`
225
+ verbatim. A machine publish attempt returns the same structured action in the
226
+ 403 error details. When run or result state is `reconciliation_required`, the
227
+ same operations URL is a required human reconciliation handoff; never infer a
228
+ URL from a status string or replay the ambiguous webhook.
229
+
178
230
  ```bash
179
231
  coil --profile prod automations create --name "Enrich leads" --webhook-url "https://hooks.example.com/enrich" --input-fields email,company_name --scope global --json
180
232
  coil --profile prod automations list --status draft --json --output-version 2
@@ -26,6 +26,11 @@ CLI discovery wrapper:
26
26
  coil agent-context --json
27
27
  ```
28
28
 
29
+ When a response requires or recommends a human web handoff, the API returns a
30
+ typed `human_action` in success `meta` or blocked error `details`, and the CLI
31
+ preserves it in JSON output. Agents must relay the exact URL and its required
32
+ or recommended state. The URL does not grant authority or imply approval.
33
+
29
34
  ## Activation, Marketplace, and Recipes
30
35
 
31
36
  | Method | Path | CLI |
@@ -95,6 +100,12 @@ Machine callers can create scrapes. Admin-gated destructive routes require a hum
95
100
 
96
101
  API-key callers create draft automations. Publishing, deletion, and policy changes require a human org admin session. Automation result links are deliberately narrow traceability records for delivery evidence; they are not a generic metadata surface.
97
102
 
103
+ Draft creation/validation, unresolved automation run/result reads, and exact
104
+ admission replays of unresolved runs may include
105
+ `meta.human_action` with the absolute `/automations` review or reconciliation
106
+ URL. Machine publication and reconciliation attempts preserve the same action
107
+ under `error.details.human_action`.
108
+
98
109
  ## Settings, Preferences, Members
99
110
 
100
111
  | Method | Path | CLI |
@@ -32,6 +32,27 @@ Use `--json --output-version 2` on list commands when possible:
32
32
  }
33
33
  ```
34
34
 
35
+ ## Human action
36
+
37
+ API success envelopes may include `meta.human_action`; blocked errors may
38
+ include the same object at `error.details.human_action`:
39
+
40
+ | Field | Type | Notes |
41
+ | --- | --- | --- |
42
+ | `kind` | `approval`/`review`/`reconciliation`/`operator_intervention` | Human handoff category. |
43
+ | `required` | boolean | Whether the human action is required to continue. |
44
+ | `label` | string | Short human-readable action description. |
45
+ | `url` | absolute HTTP(S) URL | Server-provided navigation URL. It is not a bearer token or approval capability. |
46
+ | `resource_type` | string/null | Optional Coil resource type. |
47
+ | `resource_id` | string/null | Optional resource identifier. |
48
+
49
+ The CLI preserves this action as a top-level `human_action` field for object
50
+ JSON results and V2 list envelopes. The V1 automations result list keeps its
51
+ array shape and adds `human_action` to the matching item. Text mode prints the
52
+ URL explicitly. Do not invent
53
+ one when the server does not return it. Legacy Actor Lab `approval_url` remains
54
+ available during the compatibility window.
55
+
35
56
  ## Scrape
36
57
 
37
58
  | Field | Type | Notes |
@@ -149,12 +170,18 @@ Create payload:
149
170
  | `id` | string | UUID. |
150
171
  | `automation_id` | string | Parent automation. |
151
172
  | `lead_id` | string | Processed lead. |
152
- | `status` | `pending`/`running`/`success`/`error` | Delivery state. |
173
+ | `status` | `pending`/`running`/`success`/`error`/`cancelled`/`reconciliation_required` | Delivery state. `reconciliation_required` is ambiguous provider delivery and requires a human action; it is never a retry instruction. |
153
174
  | `result` | string/null | Webhook response or summary. |
154
175
  | `error_message` | string/null | Failure detail. |
155
176
  | `created_at` | ISO string | Creation timestamp. |
156
177
  | `updated_at` | ISO string | Last status update. |
157
178
 
179
+ When automation creation or draft validation requires publication, or an
180
+ automation run/result requires reconciliation, the API returns
181
+ `meta.human_action` (or `error.details.human_action` for a blocked request).
182
+ The URL is an absolute, secret-free `/automations` navigation URL. Relay it
183
+ verbatim; opening it does not grant approval or change policy.
184
+
158
185
  ## Automation Result Link
159
186
 
160
187
  | Field | Type | Notes |