cyberfleet 0.6.0 → 0.8.0

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/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "cyberfleet",
3
- "version": "0.6.0",
3
+ "version": "0.8.0",
4
4
  "description": "Harness-agnostic, MCP-free inter-agent sessions and messaging over the filesystem",
5
5
  "keywords": [
6
6
  "agent",
@@ -38,7 +38,8 @@
38
38
  "devDependencies": {
39
39
  "@types/node": "^24.10.1",
40
40
  "commander": "^15.0.0",
41
- "cyberlegion": "^1.0.0",
41
+ "cyber-mux": "^0.12.0",
42
+ "cyberlegion": "^1.7.0",
42
43
  "tsdown": "^0.23.0",
43
44
  "tsx": "^4.22.3",
44
45
  "typescript": "^7.0.0",
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": "cyberfleet",
4
- "version": "0.6.0",
4
+ "version": "0.8.0",
5
5
  "description": "Fleet & crew personas over cyberlegion — Pod (ship's bridge), Operator (command center), Crimp (crew recruitment), and Mechanic (automaton building).",
6
6
  "author": {
7
7
  "name": "unional"
package/readme.md CHANGED
@@ -26,8 +26,8 @@ npx cyberfleet@0.0.4 <command>
26
26
  npx cyberfleet missions
27
27
  npx cyberfleet missions --agents-root <path>
28
28
 
29
- # Select/focus a ship's session (tmux pane), or print its worktree path to cd into
30
- npx cyberfleet jump <peer>
29
+ # Focus the one live unit a repo, branch, unit, or task names, or go back along the last jump
30
+ npx cyberfleet jump <target> [--repo <name>] [--branch <name>]... [--mentions <n>]
31
31
 
32
32
  # Pause a ship's mission — a cyberfleet-level status marker only (NOT the SDD pause-mission checkpoint)
33
33
  npx cyberfleet pause <peer>
@@ -112,7 +112,8 @@ Under the automatons sit two CLIs — cold, deterministic mechanism with no judg
112
112
  **`cyberfleet`** is the thin fleet layer on top, and carries only what is fleet-specific:
113
113
 
114
114
  - `missions` — the Council view: ships × mission × gate × leash, derived from SDD state
115
- - `jump` — focus a ship's pane, or print its worktree path to `cd` into
115
+ - `jump` — focus the one live unit a repo, branch, unit, or task names, or go `back`; on several
116
+ matches, a task ref, or free text it focuses nothing and returns what the `/jump` skill decides
116
117
  - `pause` — mark a ship's mission paused (a status marker only, not an SDD mission checkpoint)
117
118
  - `gate approve` — deliberately stubbed: a Council ratification can't be safely relayed through a
118
119
  CLI, so it refuses and tells you to ratify in-session
@@ -0,0 +1,37 @@
1
+ # close-sessions
2
+
3
+ A mechanical sweep that closes every agent session whose work has landed, and keeps every session
4
+ that might still have work. It has no persona and no judgment step, so it runs on a low-tier model.
5
+
6
+ ## When to use
7
+
8
+ - The Council asks to close finished agent sessions, panes, or pods, clean up idle sessions, or tidy the fleet.
9
+
10
+ Not for marking dead records only: that is `cyberlegion unit prune`, which this skill runs first. Not
11
+ for retiring a pod's record: that belongs to the pod's Captain (`cyberfleet pod retire`).
12
+
13
+ ## What it does
14
+
15
+ - Reads every unit with `cyberlegion unit who --all --format json`, and acts only on units in the
16
+ caller's current multiplexer session (records do not yet store the multiplexer session, cyberuni/cyberlegion#241, so
17
+ units whose pane `mux` matches the caller's). Other units are kept, since another session may resume them.
18
+ - Fetches pull requests once per repository and matches branches locally, instead of one `gh` call per unit.
19
+ - Classifies each unit by its worktree and its branch's pull requests. A unit is closable only when
20
+ every pull request from its branch is merged or closed and its `HEAD` is contained in a pull
21
+ request's head commit. Mechanical evidence decides first. A unit whose worktree was deleted is
22
+ closed only once its session has exited.
23
+ - Lists a landed unit with leftover uncommitted files separately, with the files, and closes it
24
+ (discarding them) only when the Council names it.
25
+ - Asks each live unit that passes the mechanical close one question by mail: "Do you have unfinished
26
+ work? Reply DONE or UNFINISHED: <reason>". A unit the multiplexer shows working is kept without
27
+ asking. UNFINISHED or no reply keeps the unit, with the reason listed. A reply never closes a unit
28
+ the evidence keeps. It never types into a pane or reads its screen, and needs no transcript or hook.
29
+ - Keeps this session, standing owners, units with no worktree, and every session in a main
30
+ checkout (Captain, Coordinator, Operator, the Council's desk).
31
+ - Shows the Council the close, ask, and keep groups, and asks about the close and ask groups
32
+ separately before closing anything.
33
+ - Closes each confirmed unit with `cyberlegion unit close <id>`, adding `--force` only for a confirmed
34
+ unit with leftover files.
35
+
36
+ This is the interim home for the sweep. The planned design moves the close into the Coordinator, at the
37
+ moment a unit's last pull request lands, and leaves this sweep for the leftovers.
@@ -0,0 +1,172 @@
1
+ ---
2
+ name: close-sessions
3
+ description: "Use this skill when the Council asks to close finished agent sessions, panes, or pods, clean up idle sessions, or tidy the fleet — it closes each unit whose worktree branch has landed and keeps everything else. Not for sweeping dead records only (`cyberlegion unit prune`), and not for retiring a Captain's pod record."
4
+ model: haiku
5
+ ---
6
+
7
+ # Close sessions
8
+
9
+ Close every agent session whose work has landed, and keep every session that might still have work.
10
+ This is a mechanical sweep: each step below is a command and a rule. The one question to a unit is
11
+ mail, and its answer can only keep the unit.
12
+
13
+ A unit is closable only when **all** of its work is visible on the host as finished: every pull
14
+ request from its branch is merged or closed, and none of its commits are missing from those pull
15
+ requests. Mechanical evidence always decides first; nothing a unit says can close one the evidence keeps.
16
+ A live unit that passes that evidence is then asked one question (step 4), because the unit is the best
17
+ judge of whether it has unfinished work. A unit whose work landed but whose worktree still has
18
+ uncommitted files is shown to the Council with those files, and closed only if the Council says so.
19
+ When a check cannot be run or reads unclear, keep the unit. Closing destroys the session, its inbox,
20
+ and its worktree, and cannot be undone. Keeping a finished unit only costs a pane.
21
+
22
+ Act only on units in the caller's current multiplexer session. Another session or multiplexer may
23
+ resume those agents, and this skill cannot know.
24
+
25
+ ## 1. Resolve `cyberlegion`
26
+
27
+ Take the first rung that resolves and whose `cyberlegion --version` is `1.6.0` or later:
28
+
29
+ 1. `cyberlegion` on `PATH`.
30
+ 2. The installed cyberlegion plugin: read `~/.claude/plugins/installed_plugins.json`, take the
31
+ `installPath` of its `cyberlegion@<marketplace>` entry, and run `node <installPath>/bin/cyberlegion.mjs`.
32
+ 3. `npx -y cyberlegion@1.6.0`.
33
+
34
+ When no rung resolves, stop and tell the Council to install the `cyberlegion` plugin.
35
+
36
+ ## 2. Read the fleet
37
+
38
+ 1. `cyberlegion unit prune` marks units whose pane is gone as exited.
39
+ 2. `cyberlegion unit whoami` gives this session's own id and handle.
40
+ 3. Name the caller's multiplexer: `herdr` when `HERDR_ENV` is set, `tmux` when `TMUX` is set.
41
+ When neither is set, stop and tell the Council.
42
+ 4. `cyberlegion unit who --all --format json` lists every unit. Each record has `id`, `handle`,
43
+ `harness`, `status`, `cwd`, a pane, and `worktree` (`root` and `branch`). The pane is under `pane`
44
+ (`mux` and `id`), or, in records written by older versions, under the legacy `tmux` key. Read it from
45
+ whichever is present. The `conversation` field is optional, and this skill does not use it.
46
+ 5. Fetch pull requests once per repository, not once per unit. For each distinct repository among the
47
+ in-scope units, run inside one of its worktrees:
48
+ `gh pr list --state all --limit 1000 --json number,state,headRefOid,headRefName`.
49
+ Match each unit's branch to `headRefName` locally. Per-unit `gh` calls made a 400-unit sweep take
50
+ about 7 minutes.
51
+ 6. When the caller's multiplexer is `herdr`, read each live in-scope pane's state once, with
52
+ `herdr pane list` (or `herdr pane get <pane id>`). It reports `agent_status` from metadata. Never read a pane's screen.
53
+
54
+ ## 3. Classify each unit
55
+
56
+ Apply these rules in order to each unit. The first rule that matches decides it.
57
+
58
+ | # | Check | Result |
59
+ | - | ----- | ------ |
60
+ | 1 | `id` is this session's own id | keep: this session |
61
+ | 2 | It is a standing owner (no pane, and no `harness` field or `harness` is `-`) | keep: standing owner |
62
+ | 3 | Its pane is absent, or its `mux` (or legacy `tmux` key) is not the caller's multiplexer | keep: out of scope |
63
+ | 4 | It has no `worktree` | keep: no worktree |
64
+ | 5 | `worktree.root` does not exist on disk, and `status` is not `exited` | keep: worktree gone, session live |
65
+ | 6 | `worktree.root` does not exist on disk, and `status` is `exited` | close: worktree gone, session ended |
66
+ | 7 | Either `git -C <root> rev-parse --git-dir` or `git -C <root> rev-parse --git-common-dir` fails | keep: cannot read worktree |
67
+ | 8 | Those two commands print the same path | keep: primary checkout |
68
+ | 9 | The repository's pull request list (step 2.5) failed to load | keep: cannot read pull requests |
69
+ | 10 | No pull request in that list has `headRefName` equal to the branch | keep: no pull request yet |
70
+ | 11 | Any matching pull request has state `OPEN` | keep: pull request open |
71
+ | 12 | `git -C <root> merge-base --is-ancestor HEAD <headRefOid>` exits non-zero for every matching pull request | keep: commits after the last pull request |
72
+ | 13 | `git -C <root> status --porcelain --untracked-files=all` prints any line other than `?? .agents/cyberlegion/config.json` | ask: work landed, leftover files |
73
+ | 14 | Anything else (every pull request is `MERGED` or `CLOSED`, `HEAD` is inside one of them, and the worktree is clean) | close: pull requests merged or closed |
74
+
75
+ Rules 5 and 6 handle a worktree that was deleted from disk. Nothing is left on disk to check, so
76
+ the session's own state decides: a live session may still be in use, and an exited one has nothing
77
+ left to lose.
78
+
79
+ Rule 3 is the scope rule. Records do not yet store the multiplexer session (cyberuni/cyberlegion#241),
80
+ so scope is the multiplexer kind. A record with no pane, or with only the legacy `tmux` key while the caller is
81
+ on herdr, is out of scope. herdr restores panes with the same ids and resumes Claude itself, so a
82
+ unit on another herdr session may come back.
83
+
84
+ Rule 8 keeps every session in a project's main checkout: a Captain, a Coordinator, the Operator, or
85
+ the Council's own desk. Those are standing seats, not spawned units, so this skill never closes one.
86
+
87
+ Rule 13 catches a unit whose work landed but left files behind: build output, scratch files, an edit it
88
+ never committed. Those are often safe to drop, but only the Council can say, so the unit is listed
89
+ with its files and closed only on the Council's word. The one line it skips is the marker cyberlegion
90
+ stamps into every worktree it spawns; `cyberlegion unit close` sets that file aside too, so it never
91
+ counts as leftover work.
92
+
93
+ Rule 12 is the unpushed-work check. It compares the local `HEAD` with each pull request's head commit
94
+ rather than with the remote branch, because the host usually deletes a branch once its pull request
95
+ merges. One pull request containing `HEAD` is enough. A head commit missing locally makes that
96
+ command exit non-zero; it counts like any other non-zero exit.
97
+
98
+ Do not retry a failed command. A failure keeps the unit, and the next sweep checks it again.
99
+
100
+ A held merge always keeps its unit, because a held pull request is still open (rule 11).
101
+
102
+ `stopped` is not `exited`. A stopped unit was kept on purpose so it can be restarted.
103
+
104
+ ## 4. Ask the live units
105
+
106
+ An **exited** unit's step-3 result stands: its session ended, so there is nobody to ask.
107
+
108
+ For each **live** unit (`status` is not `exited`) that step 3 put in **close**:
109
+
110
+ 1. When the multiplexer shows it working, keep it: **keep: working**. On herdr that is `agent_status`
111
+ `working` or `blocked`. Skip the question.
112
+ 2. Otherwise mail it one structured question on a fresh thread:
113
+
114
+ ```sh
115
+ cyberlegion mail send --to <id> --thread close-<id>-<timestamp> --subject "Closing?" \
116
+ --body "Do you have unfinished work? Reply with: cyberlegion mail send --to <caller handle> --thread <thread> --body 'DONE' (or --body 'UNFINISHED: <reason>')"
117
+ ```
118
+
119
+ `<caller handle>` is the handle from `unit whoami`, and `<thread>` is the same thread id. A reply
120
+ sent without that thread never matches the wait.
121
+
122
+ Send the questions to all such units first. Then wait on each thread with
123
+ `cyberlegion mail await --thread <thread> --timeout <ms>`, giving each the remaining share of a
124
+ total budget of about 2 minutes (120000 ms), not 2 minutes per unit. `await` prints the reply and acks
125
+ it; it exits non-zero on timeout.
126
+ 3. Read the reply:
127
+ - `DONE` → the unit stays in **close**.
128
+ - `UNFINISHED: <reason>` → **keep: unit reports unfinished work**, and show the reason.
129
+ - No reply in time, or a reply that is neither → **keep: no answer**.
130
+
131
+ A reply can only keep a unit. It never moves a unit from keep to close, and never overrides
132
+ rule 13: a live unit that passed the mechanical close but has leftover files stays in **ask**, and is
133
+ not mailed.
134
+
135
+ Never type into a pane or read its screen. The mail is the only way this skill talks to a unit.
136
+
137
+ ## 5. Confirm with the Council
138
+
139
+ Show the units in three groups: **close**, **ask**, then **keep**. Each row gives the unit id, the
140
+ handle, the branch, the result, and the pull request numbers. Handles are not unique, so the id is what
141
+ identifies a unit. Under each ask row, show what the Council needs to decide:
142
+
143
+ - for leftover files (rule 13): the lines rule 13 counted, the first 10, then how many more;
144
+
145
+ Then ask two questions, in this order:
146
+
147
+ 1. "Close these N units?" for the close group.
148
+ 2. "These M units have leftover files, listed above. Closing one deletes its worktree with those
149
+ files. Which should I close?"
150
+
151
+ The Council may drop units from either list. A yes to the first question never covers the ask group.
152
+ Close only the units the Council's answer names or covers. If an answer names a handle that more than
153
+ one candidate shares, ask which id it means. If an answer is not a yes, close nothing from that group.
154
+
155
+ ## 6. Close
156
+
157
+ For each confirmed unit, run `cyberlegion unit close <id>`. Add `--force` only for a unit that rule 13
158
+ put in ask, which discards its leftover files.
159
+
160
+ - Pass `--force` only for a rule-13 unit the Council confirmed in answer to the second question.
161
+ Never pass `--keep-worktree`.
162
+ - If a close refuses (for example, a close-group worktree that became dirty since step 3), report the
163
+ refusal and move on. Do not retry it, add `--force`, or reach around it.
164
+ - Never run `cyberfleet pod retire`. A pod's record belongs to its Captain, who retires it.
165
+ - Mail a unit only with the step-4 question. Never type into a pane or read its screen.
166
+
167
+ ## 7. Report
168
+
169
+ Lead with the count closed and the count kept (an ask-group unit the Council did not confirm counts as
170
+ kept). List the closed handles. Then list the kept units grouped by result, so the Council can see which
171
+ ones are waiting on a pull request, which reported unfinished work (with the reason), and which are
172
+ out of scope.
@@ -50,7 +50,7 @@ and never re-implements the fleet registry — it only invokes them.
50
50
  Every `cyberlegion …` command in this skill runs whichever CLI resolves below, resolved afresh in each
51
51
  session and again after any plugin reload. Never hardcode a versioned
52
52
  `~/.claude/plugins/cache/…/<version>/` path: a reload installs a new version beside the old one, and a
53
- remembered path keeps running the old CLI. The pin is `1.1.0`. Take the first rung that
53
+ remembered path keeps running the old CLI. The pin is `1.7.0`. Take the first rung that
54
54
  resolves and whose `cyberlegion --version` is at or above it; skip a rung that reports an older
55
55
  version:
56
56
 
@@ -58,11 +58,11 @@ version:
58
58
  2. The installed cyberlegion plugin: read `~/.claude/plugins/installed_plugins.json` now, take the
59
59
  `installPath` of its `cyberlegion@<marketplace>` entry, and run `node
60
60
  <installPath>/bin/cyberlegion.mjs`.
61
- 3. `npx -y cyberlegion@1.1.0`.
61
+ 3. `npx -y cyberlegion@1.7.0`.
62
62
 
63
63
  When no rung resolves, stop and report it to the Council with the
64
64
  install hint: install the `cyberlegion` plugin beside this one, or `npm install -g
65
- cyberlegion@1.1.0`.
65
+ cyberlegion@1.7.0`.
66
66
 
67
67
  ## Output
68
68
 
@@ -0,0 +1,97 @@
1
+ ---
2
+ name: jump
3
+ description: "Moves the attached view to the one pane working on a repo, a branch, a pod, a GitHub issue or pull request (`#123`, `color-map#123`), or a task described in words, and back along the last jump. Use when the Council runs /jump, or asks to go to, switch to, or show the pane working on one of these, or to go back to the pane the last jump came from (`/jump back`). Not for opening a new pane or spawning a pod, and not for messaging a unit."
4
+ argument-hint: "[repo|branch|pod|task|back|<prose>]"
5
+ ---
6
+
7
+ # Jump
8
+
9
+ Move the attached view to the one pane working on what the argument names, or back along the last
10
+ jump. `cyberfleet jump` does the matching and the focusing. This skill does the two steps the
11
+ command leaves to it: the GitHub lookup for a task ref, and matching free text to a unit.
12
+
13
+ The one rule above all others: **more than one match lists the candidates and stops.** Never pick
14
+ one, not even the newest or the most likely.
15
+
16
+ ## Running `cyberfleet`
17
+
18
+ This plugin ships the `cyberfleet` CLI, so its version always matches this skill. Run it from this
19
+ skill's directory, with no install:
20
+
21
+ ```bash
22
+ node <this skill's directory>/../../bin/cyberfleet.mjs jump <argument>
23
+ ```
24
+
25
+ Every `cyberfleet jump …` below means that command. Add `--format json` to read the result as JSON;
26
+ its `result` field names the kind.
27
+
28
+ ## 1. Run `cyberfleet jump <argument>`
29
+
30
+ Claude Code appends the argument as `ARGUMENTS: <value>`; other runtimes pass the text after the
31
+ skill name. With no argument, ask the Council where to jump and stop.
32
+
33
+ Pass the argument as it came, quoted as one value. The command prints one result:
34
+
35
+ | `result` | What it means | Next |
36
+ | -------- | ------------- | ---- |
37
+ | `jumped` | It focused one unit: its `id`, `handle`, and `pane` (`back` also gives `from`). | Step 4 |
38
+ | `candidates` | More than one unit matched. Nothing was focused. | Step 4 |
39
+ | `task` | A task ref: its `repo` name, `slug` (`<owner>/<repo>` from that checkout's `origin`), and `number`. Nothing was focused. | Step 2 |
40
+ | `prose` | No exact match: the `words`, and the live units with `id`, `handle`, `branch`, `repo`, and `brief` (a path). Nothing was focused. | Step 3 |
41
+
42
+ A non-zero exit is an error. Go to step 4 with it.
43
+
44
+ The command already applies these rules, so do not repeat them by hand: it never focuses this
45
+ session or lists it, it jumps past this session when this session was one of the matches, a bare
46
+ word must exactly equal a unit's id, handle, branch, or repo name, and `back` follows the jump
47
+ history for this pane.
48
+
49
+ ## 2. On `task`, look up the branches with `gh`
50
+
51
+ A task is a GitHub issue or pull request; one number names exactly one of them. This step needs an
52
+ authenticated `gh`. When `gh` is missing or not logged in, quote its error and stop.
53
+
54
+ 1. **Read what the number is.** `gh api repos/<slug>/issues/<number> --jq '.pull_request != null'`
55
+ prints `true` for a pull request and `false` for an issue. If it fails, quote the error and stop.
56
+ 2. **Find the branches.**
57
+ - A pull request: `gh pr view <number> -R <slug> --json headRefName --jq .headRefName`.
58
+ - An issue: `gh issue view <number> -R <slug> --json closedByPullRequestsReferences` lists the
59
+ linked pull requests. Read each one's `headRefName` as above, in that pull request's own
60
+ repository. An issue with no linked pull request has no branches.
61
+ 3. **Focus.** Run `cyberfleet jump --repo <repo> --branch <b>... --mentions <number>`, one `--branch`
62
+ per branch found, with the `repo` the `task` result gave. When no branch has a live unit,
63
+ `--mentions` falls back to a unit whose handle or brief names the number. Go to step 4 with what
64
+ it prints.
65
+
66
+ ## 3. On `prose`, match the words to one unit
67
+
68
+ Match the `words` against each listed unit's handle, branch, repo name, and brief (read the brief
69
+ file when the unit has one). A match is a unit the words plainly describe, not one they could be
70
+ stretched to fit.
71
+
72
+ - Exactly one: run `cyberfleet jump <id>` with that unit's id, and go to step 4 with what it prints.
73
+ - More than one plausible unit: list them and stop.
74
+ - None: say no live unit matches, and list the units so the Council can name one.
75
+
76
+ ## 4. Report the result
77
+
78
+ - `jumped`: one line naming the unit's handle and id and the pane, then that `/jump back` returns.
79
+ After a `back` with no unit at the pane, name the pane alone.
80
+ - `candidates`: one row per unit with its id, handle, branch, and repo name, then a line saying the
81
+ Council can narrow the argument (a repo, a branch, or an id). Do not jump.
82
+ - An error: quote the command's error verbatim with its exit code. Do not retry with other flags,
83
+ prune, or reach the pane any other way.
84
+
85
+ `/jump back` waits for this session's turn. When the Council uses `back` often, mention the tmux key
86
+ (prefix+B) that cyber-mux's `setup-tmux` binds to `focus --back`.
87
+
88
+ ## Validate
89
+
90
+ Before reporting, check the result against what the commands printed:
91
+
92
+ - A reported jump comes from a `jumped` result, and names the `id`, `handle`, and `pane` it gave.
93
+ - Every `cyberfleet jump` run after the first took an id from a result, or `--repo` with the
94
+ branches `gh` returned. None took a handle, and none was guessed.
95
+ - A candidates report has at least two rows, each a unit from the `candidates` or `prose` result,
96
+ and no `cyberfleet jump` ran after the candidates were collected.
97
+ - An error report quotes the command's error and exit code, and no other command ran after it.
@@ -150,7 +150,7 @@ the Council's words with `unit nudge --message`, never reaches for an MCP messag
150
150
  Every `cyberlegion …` command in this skill runs whichever CLI resolves below, resolved afresh in each
151
151
  session and again after any plugin reload. Never hardcode a versioned
152
152
  `~/.claude/plugins/cache/…/<version>/` path: a reload installs a new version beside the old one, and a
153
- remembered path keeps running the old CLI. The pin is `1.1.0`. Take the first rung that
153
+ remembered path keeps running the old CLI. The pin is `1.7.0`. Take the first rung that
154
154
  resolves and whose `cyberlegion --version` is at or above it; skip a rung that reports an older
155
155
  version:
156
156
 
@@ -158,11 +158,11 @@ version:
158
158
  2. The installed cyberlegion plugin: read `~/.claude/plugins/installed_plugins.json` now, take the
159
159
  `installPath` of its `cyberlegion@<marketplace>` entry, and run `node
160
160
  <installPath>/bin/cyberlegion.mjs`.
161
- 3. `npx -y cyberlegion@1.1.0`.
161
+ 3. `npx -y cyberlegion@1.7.0`.
162
162
 
163
163
  When no rung resolves, stop and report it to the Council with the
164
164
  install hint: install the `cyberlegion` plugin beside this one, or `npm install -g
165
- cyberlegion@1.1.0`.
165
+ cyberlegion@1.7.0`.
166
166
 
167
167
  ## Headless — the lifecycle loop
168
168
 
@@ -153,7 +153,7 @@ leash state itself.
153
153
  Every `cyberlegion …` command in this skill runs whichever CLI resolves below, resolved afresh in each
154
154
  session and again after any plugin reload. Never hardcode a versioned
155
155
  `~/.claude/plugins/cache/…/<version>/` path: a reload installs a new version beside the old one, and a
156
- remembered path keeps running the old CLI. The pin is `1.1.0`. Take the first rung that
156
+ remembered path keeps running the old CLI. The pin is `1.7.0`. Take the first rung that
157
157
  resolves and whose `cyberlegion --version` is at or above it; skip a rung that reports an older
158
158
  version:
159
159
 
@@ -161,11 +161,11 @@ version:
161
161
  2. The installed cyberlegion plugin: read `~/.claude/plugins/installed_plugins.json` now, take the
162
162
  `installPath` of its `cyberlegion@<marketplace>` entry, and run `node
163
163
  <installPath>/bin/cyberlegion.mjs`.
164
- 3. `npx -y cyberlegion@1.1.0`.
164
+ 3. `npx -y cyberlegion@1.7.0`.
165
165
 
166
166
  When no rung resolves, stop and report it to the Council with the
167
167
  install hint: install the `cyberlegion` plugin beside this one, or `npm install -g
168
- cyberlegion@1.1.0`.
168
+ cyberlegion@1.7.0`.
169
169
 
170
170
  ## Output
171
171