@the-open-engine-company/zeroshot 10.17.0 → 10.18.0
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 +6 -1
- package/package.json +2 -2
- package/skills/zeroshot/SKILL.md +25 -25
- package/skills/zeroshot-memory-init/SKILL.md +1 -1
package/README.md
CHANGED
|
@@ -1,6 +1,11 @@
|
|
|
1
1
|
# `@the-open-engine-company/zeroshot`
|
|
2
2
|
|
|
3
|
-
|
|
3
|
+
Zeroshot enforces your team's QA process on coding agents. You define the process as a graph:
|
|
4
|
+
which agents implement, which reviewers and tests a change must pass, where failures go for repair,
|
|
5
|
+
and where a run stops. Every task runs through it, and no agent can skip a step. It works with
|
|
6
|
+
Codex, Claude Code, GitHub Copilot and Pi.
|
|
7
|
+
|
|
8
|
+
This package is the installer for the canonical `zeroshot` executable and agent skills. The package selects the release
|
|
4
9
|
archive for the current Node platform and architecture, verifies it against that release's
|
|
5
10
|
`SHA256SUMS`, and installs the executable.
|
|
6
11
|
|
package/package.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
{
|
|
2
2
|
"name": "@the-open-engine-company/zeroshot",
|
|
3
|
-
"version": "10.
|
|
4
|
-
"description": "
|
|
3
|
+
"version": "10.18.0",
|
|
4
|
+
"description": "Enforces your team's QA process on coding agents. Define the workflow as a graph of implementers, reviewers, tests and repair loops; every task runs through it and no agent can skip a step. Works with Codex, Claude Code, GitHub Copilot and Pi.",
|
|
5
5
|
"keywords": [
|
|
6
6
|
"claude-code",
|
|
7
7
|
"codex",
|
package/skills/zeroshot/SKILL.md
CHANGED
|
@@ -1,6 +1,6 @@
|
|
|
1
1
|
---
|
|
2
2
|
name: zeroshot
|
|
3
|
-
description: Use Zeroshot to prepare, run, observe, or troubleshoot explicit multi-agent software work locally or on Zeroshot
|
|
3
|
+
description: Use Zeroshot to prepare, run, observe, or troubleshoot explicit multi-agent software work locally or on Zeroshot Platform. Apply when the user names Zeroshot, including 'zeroshot this' or 'zeroshot it' (this orchestrator, not one-shot prompting), or asks about its CLI; do not route ordinary local work into Zeroshot without that intent.
|
|
4
4
|
---
|
|
5
5
|
|
|
6
6
|
# Zeroshot
|
|
@@ -26,7 +26,7 @@ Suggest separate tasks only for independently useful outcomes or real prerequisi
|
|
|
26
26
|
|
|
27
27
|
Never start a Zeroshot run merely because the skill is available. The user must ask to use Zeroshot.
|
|
28
28
|
A worker already executing inside a Zeroshot run should perform its assigned work directly, not start
|
|
29
|
-
another run. A failed
|
|
29
|
+
another run. A failed Platform run does not authorize silently implementing the task locally.
|
|
30
30
|
|
|
31
31
|
## Choose the target
|
|
32
32
|
|
|
@@ -34,10 +34,10 @@ another run. A failed Cloud run does not authorize silently implementing the tas
|
|
|
34
34
|
Git worktrees use their root. File arguments stay relative to the invoking folder. Runs without
|
|
35
35
|
Git delivery need no repository. Keep workspace and active Zeroshot state disjoint; use an external
|
|
36
36
|
`ZEROSHOT_STATE_DIR` when necessary.
|
|
37
|
-
- Zeroshot
|
|
38
|
-
`--target
|
|
37
|
+
- Zeroshot Platform is built in. Sign in once with `zeroshot target login platform`, then use
|
|
38
|
+
`--target platform` explicitly on Platform commands. Platform setup (organizations, GitHub App,
|
|
39
39
|
connections) is documented at https://cloud.zeroshot.sh/docs.
|
|
40
|
-
- Do not add or set up the built-in
|
|
40
|
+
- Do not add or set up the built-in Platform target. Named targets store endpoints and login identity,
|
|
41
41
|
not repository configuration.
|
|
42
42
|
|
|
43
43
|
## Reuse profiles and connections
|
|
@@ -49,11 +49,11 @@ when a run names no profile, graph, or runtime:
|
|
|
49
49
|
```console
|
|
50
50
|
zeroshot profile list
|
|
51
51
|
zeroshot profile show NAME
|
|
52
|
-
zeroshot profile list --target
|
|
53
|
-
zeroshot profile list --target
|
|
54
|
-
zeroshot profile show NAME --target
|
|
55
|
-
zeroshot connection list --target
|
|
56
|
-
zeroshot connection list --target
|
|
52
|
+
zeroshot profile list --target platform --scope user
|
|
53
|
+
zeroshot profile list --target platform --scope org
|
|
54
|
+
zeroshot profile show NAME --target platform --scope org
|
|
55
|
+
zeroshot connection list --target platform --scope user
|
|
56
|
+
zeroshot connection list --target platform --scope org
|
|
57
57
|
```
|
|
58
58
|
|
|
59
59
|
Inspect saved profiles to see their delivery behavior. New `software-change` templates default to
|
|
@@ -62,7 +62,7 @@ or `--ship` when merge delivery is intended. Custom graphs and saved profiles ow
|
|
|
62
62
|
|
|
63
63
|
## Run locally without a profile
|
|
64
64
|
|
|
65
|
-
When no saved profile fits and the user has not asked for
|
|
65
|
+
When no saved profile fits and the user has not asked for Platform, run the built-in
|
|
66
66
|
`software-change` graph locally. Its default publishes a PR. For local-only work, use
|
|
67
67
|
`--delivery none` as shown below; that mode needs only the harness's native login.
|
|
68
68
|
|
|
@@ -116,7 +116,7 @@ zeroshot run --title "Review PR 123" --template pr-review \
|
|
|
116
116
|
|
|
117
117
|
Use `issue-review` for an issue URL; its basic issue workflow is separate. PR intake captures linked
|
|
118
118
|
issue text before execution. Local reviews require clean tracked files and, for a PR, a checkout already at its head. Comment
|
|
119
|
-
publication uses the declared `github.GH_TOKEN` connection. Use `--target
|
|
119
|
+
publication uses the declared `github.GH_TOKEN` connection. Use `--target platform` for hosted source
|
|
120
120
|
preparation and credential resolution. With a saved review profile, use `--profile` instead of the
|
|
121
121
|
template/runtime/delivery flags. Validate first with `--validate-only`; retry uncertain submission
|
|
122
122
|
using the same key and configuration. A successful review may report defects; success confirms
|
|
@@ -150,7 +150,7 @@ one from scratch:
|
|
|
150
150
|
## Prepare the execution environment
|
|
151
151
|
|
|
152
152
|
Environments belong to a run, independently of its profile. A `local:NAME` profile can run on a
|
|
153
|
-
Docker target with its own environment. For Docker or
|
|
153
|
+
Docker target with its own environment. For Docker or Platform, pass
|
|
154
154
|
`--environment environment.json` with a flat definition such as:
|
|
155
155
|
|
|
156
156
|
```json
|
|
@@ -161,8 +161,8 @@ Docker target with its own environment. For Docker or Cloud, pass
|
|
|
161
161
|
}
|
|
162
162
|
```
|
|
163
163
|
|
|
164
|
-
|
|
165
|
-
Base. `--no-environment` bypasses defaults. Choose saved environments through
|
|
164
|
+
Platform resolves an omitted environment from the repository's User default, then Org default, then
|
|
165
|
+
Base. `--no-environment` bypasses defaults. Choose saved environments through Platform's UI or API;
|
|
166
166
|
the CLI accepts a concrete JSON file. Local runs use the invoking machine and reject setup/startup
|
|
167
167
|
and hook connections.
|
|
168
168
|
|
|
@@ -213,23 +213,23 @@ submission key for safe retries:
|
|
|
213
213
|
|
|
214
214
|
```console
|
|
215
215
|
zeroshot run --title "TASK" --profile org:NAME --input input.json \
|
|
216
|
-
--target
|
|
216
|
+
--target platform --submission-key KEY --validate-only
|
|
217
217
|
|
|
218
218
|
zeroshot run --title "TASK" --profile org:NAME --input input.json \
|
|
219
|
-
--target
|
|
219
|
+
--target platform --submission-key KEY
|
|
220
220
|
```
|
|
221
221
|
|
|
222
|
-
`--validate-only` materializes and validates without starting a run. A named
|
|
223
|
-
still be fetched from the target first, so this command requires
|
|
222
|
+
`--validate-only` materializes and validates without starting a run. A named Platform profile must
|
|
223
|
+
still be fetched from the target first, so this command requires Platform access. It does not prove
|
|
224
224
|
connection availability.
|
|
225
225
|
|
|
226
226
|
For named targets, `--cwd` selects the client Git checkout. Zeroshot prefers its attached upstream,
|
|
227
|
-
then a unique eligible HTTPS/SSH remote. Docker and
|
|
227
|
+
then a unique eligible HTTPS/SSH remote. Docker and Platform still require an exact committed Git
|
|
228
228
|
source. Detached worktrees require repository and branch overrides. `--repository` accepts GitHub
|
|
229
229
|
`owner/name` or a credential-free HTTPS/SSH clone URL. Override source only when needed:
|
|
230
230
|
|
|
231
231
|
```console
|
|
232
|
-
--target
|
|
232
|
+
--target platform --repository OWNER/REPOSITORY --branch BRANCH --revision SHA
|
|
233
233
|
```
|
|
234
234
|
|
|
235
235
|
Git push uses shared transport and needs no GitHub API. PR and merge require a registered forge;
|
|
@@ -246,13 +246,13 @@ profile, and delivery intent reported for the run.
|
|
|
246
246
|
Keep the run ID and target:
|
|
247
247
|
|
|
248
248
|
```console
|
|
249
|
-
zeroshot status RUN_ID --target
|
|
250
|
-
zeroshot watch RUN_ID --target
|
|
251
|
-
zeroshot logs RUN_ID --target
|
|
249
|
+
zeroshot status RUN_ID --target platform
|
|
250
|
+
zeroshot watch RUN_ID --target platform
|
|
251
|
+
zeroshot logs RUN_ID --target platform
|
|
252
252
|
```
|
|
253
253
|
|
|
254
254
|
A foreground run streams events. Ctrl-C detaches observation without stopping the run. Use
|
|
255
|
-
`zeroshot force-stop RUN_ID --target
|
|
255
|
+
`zeroshot force-stop RUN_ID --target platform` only when the user explicitly intends to stop it.
|
|
256
256
|
|
|
257
257
|
For command errors, check `zeroshot --version` and the relevant `--help`; do not revive removed
|
|
258
258
|
commands such as `target setup`. If login cannot retain credentials, fix the credential store before
|
|
@@ -84,7 +84,7 @@ For an invalid response, correct only the named task using its existing contract
|
|
|
84
84
|
failure persists, stop and report the error and session directory rather than retry indefinitely.
|
|
85
85
|
Report exclusions and incomplete work alongside the command's publication/defer counts.
|
|
86
86
|
|
|
87
|
-
This workflow reads local native Rust run history, not
|
|
87
|
+
This workflow reads local native Rust run history, not Platform or Docker target history. Never read
|
|
88
88
|
controller bootstrap files, provider homes or credentials to recover evidence. Do not fetch code or
|
|
89
89
|
switch branches to interpret “newest”; use the recorded checkouts. Current maintenance citations must
|
|
90
90
|
refer to tracked or nonignored source, which the CLI fingerprints before publication.
|