@pulsarinteractive/geniustasker 0.1.0-beta.16 → 0.1.0-beta.18
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/CHANGELOG.md +12 -0
- package/README.md +1 -1
- package/docs/CHATGPT.md +18 -0
- package/docs/CLI.md +28 -1
- package/docs/DOCUMENTATION_SYNC.md +8 -0
- package/docs/INTEGRATIONS.md +4 -4
- package/docs/MEMORY.md +90 -8
- package/docs/TASK_CHECKLISTS.md +20 -0
- package/docs/WORKFLOWS.md +8 -0
- package/examples/codex.toml +1 -1
- package/examples/mcp.json +1 -1
- package/mcp.json +1 -1
- package/package.json +1 -1
- package/plugin.json +1 -1
- package/remote/tool-catalog.json +496 -7
- package/remote/workflows.json +202 -6
- package/skills/geniustasker/SKILL.md +11 -0
- package/skills/geniustasker/references/importance.md +29 -0
- package/skills/geniustasker-documentation/SKILL.md +5 -0
- package/skills/geniustasker-feedback/SKILL.md +2 -0
- package/skills/geniustasker-memory/SKILL.md +31 -5
- package/skills/geniustasker-planning/SKILL.md +7 -0
- package/src/archive.d.mts +1 -0
- package/src/archive.mjs +96 -0
- package/src/cli.mjs +89 -1
- package/src/client.mjs +6 -0
- package/src/command-catalog.mjs +24 -1
- package/src/command-input-schemas.mjs +253 -0
- package/src/commands.mjs +11 -1
- package/src/documentation-sync.mjs +23 -2
- package/src/mcp.mjs +46 -5
- package/src/memory-editorial.d.mts +5 -0
- package/src/memory-editorial.mjs +216 -0
- package/src/memory-history.mjs +45 -0
- package/src/memory-read-pages.d.mts +4 -0
- package/src/memory-read-pages.mjs +221 -0
- package/src/presentation.d.mts +1 -0
- package/src/presentation.mjs +16 -0
- package/src/resources.mjs +2 -2
package/CHANGELOG.md
CHANGED
|
@@ -1,5 +1,17 @@
|
|
|
1
1
|
# Changelog
|
|
2
2
|
|
|
3
|
+
## 0.1.0-beta.18 — 2026-11-19
|
|
4
|
+
|
|
5
|
+
- Review verified project changes against the owner's documentation policy and the selected note revision. Offer first notes individually, preserve each section's language and style, and preview optional product, game, music or free-form pages.
|
|
6
|
+
- Read selected page outlines, blocks and retained revisions in metered batches of up to five. Inspect revision history, compare changes and restore an explicitly chosen version.
|
|
7
|
+
- Query one project's archive by text, dates, status, impact or assignee, then read an archived task and its comments with explicit historical access. Existing grants remain unchanged.
|
|
8
|
+
|
|
9
|
+
## 0.1.0-beta.17
|
|
10
|
+
|
|
11
|
+
- Discover support reporting through feedback capabilities, including when ordinary project commands omit it.
|
|
12
|
+
- Configure completion alerts independently from task impact. Respect human-owned choices and nominate exceptional subtasks explicitly.
|
|
13
|
+
- Read one task’s description and its checklist by task ID. Add steps to a chosen section, including while holding an agent run.
|
|
14
|
+
|
|
3
15
|
## 0.1.0-beta.16
|
|
4
16
|
|
|
5
17
|
- Document how an agent starts a New task after claiming its run, verifies InProgress on the task, and explicitly resumes previously paused work.
|
package/README.md
CHANGED
|
@@ -55,7 +55,7 @@ For a stdio MCP host, use this after signing in:
|
|
|
55
55
|
"mcpServers": {
|
|
56
56
|
"geniustasker": {
|
|
57
57
|
"command": "npx",
|
|
58
|
-
"args": ["-y", "@pulsarinteractive/geniustasker@0.1.0-beta.
|
|
58
|
+
"args": ["-y", "@pulsarinteractive/geniustasker@0.1.0-beta.18", "mcp", "--context", "work"]
|
|
59
59
|
}
|
|
60
60
|
}
|
|
61
61
|
}
|
package/docs/CHATGPT.md
CHANGED
|
@@ -107,3 +107,21 @@ loads the canonical change. When reconnecting after a network interruption, the
|
|
|
107
107
|
application catches up through its existing sync system. A server receipt and a
|
|
108
108
|
screen update are different observations: the assistant should verify the changed
|
|
109
109
|
record, and should not claim to have seen your screen unless it actually did.
|
|
110
|
+
|
|
111
|
+
## If ChatGPT cannot find bug reports or feature requests
|
|
112
|
+
|
|
113
|
+
Select **@GeniusTasker** and ask whether this connection can report a bug or suggest a
|
|
114
|
+
feature. The live check is `tasker_feedback_capabilities`. Product feedback
|
|
115
|
+
uses dedicated tools outside the list of project or training commands.
|
|
116
|
+
|
|
117
|
+
1. If that tool is missing, open **Plugins → GeniusTasker → Settings → Manage app → Refresh tools**, then select the plugin in a new chat.
|
|
118
|
+
2. If the tool returns `canReportBug: false`, review the selected account and its personal `feedback:write` grant. Reconnect in ChatGPT if you want to authorize that access.
|
|
119
|
+
3. If `canReportBug` is true and `canRequestFeature` is false, this account can report bugs. Feature requests require a paid plan.
|
|
120
|
+
4. Ask to prepare a report. Review its contents and selected attachments, then approve sending. The returned receipt identifies the submission.
|
|
121
|
+
|
|
122
|
+
When a verified feature changes, ask the connector to check the linked project's
|
|
123
|
+
editorial policy and its existing Memory. The `memory-editorial` workflow
|
|
124
|
+
describes when it may edit, offer a draft or leave the result in its ticket.
|
|
125
|
+
For comparing notes, use `tasker_memory_read_pages` with specific authorized
|
|
126
|
+
page and block IDs. The `memory-reading`, `memory-history` and
|
|
127
|
+
`memory-templates` workflows apply directly in hosted ChatGPT.
|
package/docs/CLI.md
CHANGED
|
@@ -131,7 +131,7 @@ launching stdio servers with this entry (adjust environment/context as needed):
|
|
|
131
131
|
"mcpServers": {
|
|
132
132
|
"geniustasker": {
|
|
133
133
|
"command": "npx",
|
|
134
|
-
"args": ["-y", "@pulsarinteractive/geniustasker@0.1.0-beta.
|
|
134
|
+
"args": ["-y", "@pulsarinteractive/geniustasker@0.1.0-beta.18", "mcp", "--context", "work"]
|
|
135
135
|
}
|
|
136
136
|
}
|
|
137
137
|
}
|
|
@@ -638,3 +638,30 @@ MCP tasker_resources/tasker_read_tasks.
|
|
|
638
638
|
Inspect exact schemas with `commands describe --kind checklist.create` (also
|
|
639
639
|
update/move/remove or run.checklist under a live claim). Checking is an explicit
|
|
640
640
|
boolean, never a toggle or automatic task completion.
|
|
641
|
+
|
|
642
|
+
## Search archived tasks
|
|
643
|
+
|
|
644
|
+
Use `archive query --file query.json` or `tasker_archive_query` to retrieve
|
|
645
|
+
one page for a project. Both `tasks:read` and `history:read` must be granted
|
|
646
|
+
on that project. Existing connections keep their prior grants until the owner
|
|
647
|
+
consents to the additional permission.
|
|
648
|
+
|
|
649
|
+
```json
|
|
650
|
+
{
|
|
651
|
+
"scope": "project:example",
|
|
652
|
+
"query": "release",
|
|
653
|
+
"importance": ["Important", "Critical"],
|
|
654
|
+
"archivedFrom": "2026-10-01T00:00:00Z",
|
|
655
|
+
"limit": 20
|
|
656
|
+
}
|
|
657
|
+
```
|
|
658
|
+
|
|
659
|
+
Also supported are `archivedUntil`, `statuses`, `assignees`, `sort`,
|
|
660
|
+
`sharedFilterId` and `column`. Read the next page by supplying the returned
|
|
661
|
+
`next` as `continuation` together with the unchanged query. A page can be
|
|
662
|
+
empty while a continuation remains. Fetch an item with
|
|
663
|
+
`archive detail --scope project:ID --id TASK` and load comments separately
|
|
664
|
+
with `archive comments --scope project:ID --id TASK`. The connector offers
|
|
665
|
+
`tasker_archive_detail` and `tasker_archive_comments`. A restored, expired or
|
|
666
|
+
inaccessible task may be absent from the archive. Reconnect with explicit
|
|
667
|
+
historical access when needed.
|
|
@@ -95,6 +95,14 @@ reset therefore does not require downloading the same first pages repeatedly.
|
|
|
95
95
|
Changed or inaccessible entries are discarded. The temporary cache is bounded
|
|
96
96
|
to twice the selected byte budget and is removed after a successful sync.
|
|
97
97
|
|
|
98
|
+
The CLI and local MCP return per-scope progress on a rate limit. For example,
|
|
99
|
+
`preparedPages: 137, selectedPages: 500, phase: "pages"` means that 137 page
|
|
100
|
+
bodies were prepared before the interruption. During inventory the total is
|
|
101
|
+
unknown (`null`). Report the remaining scope of work, respect the returned retry
|
|
102
|
+
time, then repeat the same command and destination. On an interrupted first
|
|
103
|
+
sync, pass the configuration file again. The final authorization check runs
|
|
104
|
+
before the complete snapshot becomes visible.
|
|
105
|
+
|
|
98
106
|
Use `memory tree --scope memory:ID` to inspect a bounded folder. Continue with
|
|
99
107
|
`--after`, `--epoch` and `--checkpoint`. Use `memory markdown --scope memory:ID
|
|
100
108
|
--page PAGE_ID` for one page, optionally with its expected `--revision`.
|
package/docs/INTEGRATIONS.md
CHANGED
|
@@ -16,8 +16,8 @@ environment or replace a connection just to get around an access refusal.
|
|
|
16
16
|
For example, to connect a production context to a direct Codex MCP server:
|
|
17
17
|
|
|
18
18
|
```sh
|
|
19
|
-
npx -y @pulsarinteractive/geniustasker@0.1.0-beta.
|
|
20
|
-
codex mcp add geniustasker-production -- npx -y @pulsarinteractive/geniustasker@0.1.0-beta.
|
|
19
|
+
npx -y @pulsarinteractive/geniustasker@0.1.0-beta.18 auth login --environment production --context tasker-beta
|
|
20
|
+
codex mcp add geniustasker-production -- npx -y @pulsarinteractive/geniustasker@0.1.0-beta.18 mcp --environment production --context tasker-beta
|
|
21
21
|
```
|
|
22
22
|
|
|
23
23
|
Use either this configuration or the development example below for the intended
|
|
@@ -27,8 +27,8 @@ before working. Adding the production example does not retarget a bundled plugin
|
|
|
27
27
|
## Codex: direct MCP
|
|
28
28
|
|
|
29
29
|
```sh
|
|
30
|
-
npx -y @pulsarinteractive/geniustasker@0.1.0-beta.
|
|
31
|
-
codex mcp add geniustasker -- npx -y @pulsarinteractive/geniustasker@0.1.0-beta.
|
|
30
|
+
npx -y @pulsarinteractive/geniustasker@0.1.0-beta.18 auth login --context work
|
|
31
|
+
codex mcp add geniustasker -- npx -y @pulsarinteractive/geniustasker@0.1.0-beta.18 mcp --context work
|
|
32
32
|
```
|
|
33
33
|
|
|
34
34
|
Restart the host as needed, then inspect its MCP list. Ask the agent to read its
|
package/docs/MEMORY.md
CHANGED
|
@@ -1,13 +1,12 @@
|
|
|
1
1
|
# Memory and Graph
|
|
2
2
|
|
|
3
|
-
|
|
4
|
-
|
|
5
|
-
|
|
6
|
-
|
|
7
|
-
|
|
8
|
-
|
|
9
|
-
|
|
10
|
-
package installation alone does not activate a server release.
|
|
3
|
+
Delegated access to Memory documents, history, assets and Graph through the CLI,
|
|
4
|
+
local MCP and ChatGPT requires a current Starter or Pro account and explicit
|
|
5
|
+
rights to the selected resource. A linked project and a paid space owner confer
|
|
6
|
+
no access on another account. Read live capabilities before selecting sections,
|
|
7
|
+
history or restore. On `memory_premium_required` or
|
|
8
|
+
`memory_graph_premium_required`, end that delegated read. The ordinary app and
|
|
9
|
+
account export use their own existing permissions after a downgrade.
|
|
11
10
|
|
|
12
11
|
Memory stores structured documents independently of projects. Graph connects
|
|
13
12
|
nodes and links them to a page or a stable document block. These commands require
|
|
@@ -45,3 +44,86 @@ so people can see and repair the unavailable attachment. No credentials, URLs
|
|
|
45
44
|
with public access tokens or file bytes are embedded in documents.
|
|
46
45
|
|
|
47
46
|
See the [CLI reference](CLI.md) for document blocks, page roles, node types, images and detailed command examples.
|
|
47
|
+
|
|
48
|
+
## Maintain a project note after a feature changes
|
|
49
|
+
|
|
50
|
+
At the start of an actual work session, inspect the project's linked library,
|
|
51
|
+
destination and editorial policy. After verifying a lasting change, search for
|
|
52
|
+
the existing feature guide and read its relevant section. An implemented fix, a
|
|
53
|
+
release and a proposal describe different states. Use the verified state in the
|
|
54
|
+
note and keep routine task progress in its ticket. When no suitable note exists,
|
|
55
|
+
offer a specific first page for an identified reader and purpose. Respect a
|
|
56
|
+
previous decision to defer or decline.
|
|
57
|
+
|
|
58
|
+
The command `geniustasker workflow show --topic memory-editorial` and MCP tool
|
|
59
|
+
`tasker_workflow` publish the same allowed editing procedure. The
|
|
60
|
+
`memory-templates` topic provides optional product, game, music and free-form
|
|
61
|
+
pages with English and French starting headings. Preview the selected page,
|
|
62
|
+
adapt the wording to the destination and populate it from authorized evidence.
|
|
63
|
+
|
|
64
|
+
## Read several notes
|
|
65
|
+
|
|
66
|
+
Save one JSON selection, then run `geniustasker memory read-pages --file selection.json`.
|
|
67
|
+
The local MCP and ChatGPT use the same JSON in `tasker_memory_read_pages`.
|
|
68
|
+
|
|
69
|
+
```json
|
|
70
|
+
{
|
|
71
|
+
"scope": "memory:example",
|
|
72
|
+
"pages": [
|
|
73
|
+
{ "pageId": "release-guide", "mode": "outline" },
|
|
74
|
+
{ "pageId": "offline-workouts", "mode": "blocks", "blockIds": ["starting-offline"] }
|
|
75
|
+
]
|
|
76
|
+
}
|
|
77
|
+
```
|
|
78
|
+
|
|
79
|
+
Use IDs returned by the authorized index and outline. Each selected page
|
|
80
|
+
uses one ordinary account read. A batch accepts five pages and runs at most two
|
|
81
|
+
reads at once. The logical response is limited to 64 KiB of UTF-8 JSON.
|
|
82
|
+
`complete` counts fully consumed selectors, including unchanged selectors.
|
|
83
|
+
A complete outline covers headings and previews only. Each partial page returns
|
|
84
|
+
its revision and a `nextCursor`. Reuse the original selector, supply that cursor
|
|
85
|
+
as `cursor`, and include the returned revision, scope epoch and checkpoint on
|
|
86
|
+
the next call. Use the stable block ID and UTF-16 offset to reassemble split text.
|
|
87
|
+
After a quota interruption save the unread IDs, wait for the reported reset,
|
|
88
|
+
and resume from the last verified page. No background reads are scheduled.
|
|
89
|
+
|
|
90
|
+
Compare ten chosen notes in two groups of five when each fits. Search and read
|
|
91
|
+
matching sections for a focused question about a 500-page library. For an
|
|
92
|
+
explicit full review, enumerate authorized IDs and revisions into a resumable
|
|
93
|
+
manifest, count complete and outstanding pages, then process selected blocks in
|
|
94
|
+
bounded batches. An expired checkpoint requires a new inventory and fresh
|
|
95
|
+
permission checks. A missing or forbidden page stays unavailable.
|
|
96
|
+
|
|
97
|
+
## Read and restore revisions
|
|
98
|
+
|
|
99
|
+
`memory history --scope memory:ID --page PAGE` lists retained revisions.
|
|
100
|
+
Repeat with `--before-revision` from `nextBeforeRevision` for older entries.
|
|
101
|
+
`memory diff --scope memory:ID --page PAGE --from REV --to REV` compares
|
|
102
|
+
retained revisions. To read retained content, include its revision as `atRevision` on the original outline or block selector. Reuse the same historical selector when continuing. Current membership and page rights always apply. The corresponding tools are `tasker_memory_history` and
|
|
103
|
+
`tasker_memory_diff`. When explicitly authorized, prepare the chosen
|
|
104
|
+
`scope`, `pageId`, `revision`, current `expectedRevision`, `epoch`,
|
|
105
|
+
new UUID `operationId` and millisecond `issuedAt` once. Submit the exact object
|
|
106
|
+
to `tasker_memory_restore` or `memory restore --file restore.json`. Preserve it
|
|
107
|
+
for retries and inspect the resulting note and revision after success.
|
|
108
|
+
|
|
109
|
+
Use `geniustasker memory template --family game --language fr` to list
|
|
110
|
+
independent game pages. Add `--page mechanic` to preview that page in French.
|
|
111
|
+
The preview contains possible headings and performs no write. The connector
|
|
112
|
+
offers the same selections through `tasker_memory_template`.
|
|
113
|
+
|
|
114
|
+
For a verified milestone, `memory review --file review.json` or
|
|
115
|
+
`tasker_memory_review` reads the project's current policy and checks the selected
|
|
116
|
+
candidate revision. Supply an editorial domain, the type of change, references
|
|
117
|
+
to reviewed evidence and candidates from a scoped search. The result explains
|
|
118
|
+
whether to leave the information in the task, clarify it, suggest a note or prepare
|
|
119
|
+
a small update. A read-only review cannot execute or authorize that update.
|
|
120
|
+
|
|
121
|
+
```json
|
|
122
|
+
{
|
|
123
|
+
"scope": "project:example",
|
|
124
|
+
"domain": "usage",
|
|
125
|
+
"change": "feature",
|
|
126
|
+
"sources": [{ "kind": "tested-release", "reference": "reviewed-release-42" }],
|
|
127
|
+
"candidates": [{ "pageId": "offline-workouts", "revision": "21" }]
|
|
128
|
+
}
|
|
129
|
+
```
|
package/docs/TASK_CHECKLISTS.md
CHANGED
|
@@ -106,3 +106,23 @@ does not mark the task Done. Editing, checking, moving or clearing a checklist
|
|
|
106
106
|
does not start, resume, unlock, reopen or move Paused work to New. Leave actual
|
|
107
107
|
progress and blockers intact; complete the task separately only after verifying
|
|
108
108
|
its intended outcome and acceptance criteria.
|
|
109
|
+
|
|
110
|
+
### Sections in the task description
|
|
111
|
+
|
|
112
|
+
When the service advertises `sectionId` on `checklist.create`, read the
|
|
113
|
+
task document with `tasker_resources` or CLI resources using collection
|
|
114
|
+
`documents` and the same `taskId`. An ordered Quill insert such as
|
|
115
|
+
`{"insert":{"task-checklist":"<section UUID>"}}` marks a section. Read the
|
|
116
|
+
canonical checklist for that task, then match items by `sectionId` and sort
|
|
117
|
+
each section by `order`. Items with a missing or null section belong to the
|
|
118
|
+
default list. Preserve any items whose section is absent from an older
|
|
119
|
+
document, and surface them as a separate list.
|
|
120
|
+
|
|
121
|
+
Add to the chosen section with `checklist.create` and its exact section UUID.
|
|
122
|
+
The same field is accepted inside a claimed run’s `run.checklist.change`.
|
|
123
|
+
Omit it to use the default list. Rename, check and remove still target the
|
|
124
|
+
item ID and revision. A move swaps adjacent items in the same section and
|
|
125
|
+
requires both revisions. Reordering a whole section changes its position in
|
|
126
|
+
the document, with the current document version. A populated section cannot
|
|
127
|
+
be removed by deleting its embed. Keep the saved steps unless the user asks
|
|
128
|
+
for their removal explicitly. Task counts cover every section.
|
package/docs/WORKFLOWS.md
CHANGED
|
@@ -49,3 +49,11 @@ Read [Task protocols](../skills/geniustasker/references/protocols.md) for exact
|
|
|
49
49
|
For real agent claims, checkpoints, scoped task patches and completion, read
|
|
50
50
|
[Agent runs](../skills/geniustasker/references/agent-runs.md). A recorded run does
|
|
51
51
|
not start or monitor an external agent automatically.
|
|
52
|
+
|
|
53
|
+
## Memory maintenance and reading
|
|
54
|
+
|
|
55
|
+
- `memory-editorial` evaluates verified project changes, document ownership, section language, writing style, proposals and conflicts.
|
|
56
|
+
- `memory-reading` covers selected sections, resumable multi-page comparisons and explicitly exhaustive inventories.
|
|
57
|
+
- `archives` covers bounded, permission-specific task and comment lookups for historical work.
|
|
58
|
+
- `memory-history` covers revision metadata, scoped diffs and explicit restores.
|
|
59
|
+
- `memory-templates` previews an individually selected product, game, music or free-form page.
|
package/examples/codex.toml
CHANGED
package/examples/mcp.json
CHANGED
package/mcp.json
CHANGED
package/package.json
CHANGED
package/plugin.json
CHANGED
|
@@ -1,7 +1,7 @@
|
|
|
1
1
|
{
|
|
2
2
|
"$schema": "https://agent-plugins.org/schemas/1.0.0/plugin.schema.json",
|
|
3
3
|
"name": "geniustasker",
|
|
4
|
-
"version": "0.1.0-beta.
|
|
4
|
+
"version": "0.1.0-beta.18",
|
|
5
5
|
"description": "Track project work and independent product memories with your own agents.",
|
|
6
6
|
"author": {
|
|
7
7
|
"name": "Pulsar Interactive"
|