@sjawhar/opencode-legion-envoy 6.2.0 → 6.3.0
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
|
@@ -230,9 +230,10 @@ review and the merge-gate sequence by the daemon's table.
|
|
|
230
230
|
|
|
231
231
|
Retro is mandatory for every issue that passed review, before merge, and the daemon runs it: when
|
|
232
232
|
the reviewer's approval ends the review round, it moves the issue to `retro` and starts the
|
|
233
|
-
implementer on it,
|
|
234
|
-
|
|
235
|
-
|
|
233
|
+
implementer on it, in the same agent's session, which stayed live after its phase ended (every
|
|
234
|
+
started role stays live until its issue closes). You start nothing for it. Never `envoy_publish` to
|
|
235
|
+
the implementer's role topic to start retro: the daemon's start is what records which role owns
|
|
236
|
+
the task.
|
|
236
237
|
|
|
237
238
|
Wait for the implementer to report its durable retro result. Retro output is
|
|
238
239
|
`docs/solutions/`, the PR body content the repository's instructions derive from the pull
|
|
@@ -68,10 +68,10 @@ into a spec yourself: the architect decides whether the human must answer it and
|
|
|
68
68
|
since a new version of an approved root spec closes the tree's design gate. A standalone to-do
|
|
69
69
|
only a human can do is a `dispatch_ask`, and its replies return to your own session.
|
|
70
70
|
|
|
71
|
-
Because the same agent
|
|
72
|
-
assignment across your lifetime:
|
|
73
|
-
event (a review round, a red check) starts your role again
|
|
74
|
-
|
|
71
|
+
Because the same agent works its role until the issue closes, you may receive more than one
|
|
72
|
+
assignment across your lifetime: your session stays live after your phase ends, and when a later
|
|
73
|
+
event (a review round, a red check) starts your role again the new prompt arrives in this same
|
|
74
|
+
session. Treat it as a continuation — re-read the current issue and your own prior handoff, since
|
|
75
75
|
time has passed — never as a fresh identity.
|
|
76
76
|
|
|
77
77
|
## Deployment instructions
|
|
@@ -93,9 +93,9 @@ implementer's production check after the merge
|
|
|
93
93
|
Reach any live role on this issue the same way you reach the architect: `envoy_publish` to
|
|
94
94
|
`notifications.role.` followed by that role's encoded token. Use it when you need context an
|
|
95
95
|
earlier phase has that its handoff doesn't cover — ask the planner why a constraint was
|
|
96
|
-
scoped that way, ask the implementer what a commit actually did.
|
|
97
|
-
|
|
98
|
-
|
|
96
|
+
scoped that way, ask the implementer what a commit actually did. A role that finished its phase
|
|
97
|
+
stays live and answers until its issue leaves the workflow. If a publish is rejected with 404,
|
|
98
|
+
read that role's committed handoff.
|
|
99
99
|
|
|
100
100
|
## Workspace and handoff precedence
|
|
101
101
|
|
|
@@ -111,10 +111,14 @@ Never rely on the inherited cwd. Every later repository shell command **MUST** b
|
|
|
111
111
|
native filesystem tool paths **MUST** be absolute under that workspace. Do not create an
|
|
112
112
|
isolated worktree, change the workspace topology, or mix another issue's work into it.
|
|
113
113
|
Concurrent issues have disjoint workspaces; only the currently active phase mutates this
|
|
114
|
-
one. After you complete, treat `$LEGION_WORKSPACE` as read-only
|
|
115
|
-
|
|
116
|
-
`jj -R "$LEGION_WORKSPACE" new`, or touch tracked files
|
|
117
|
-
(and, for the implementer, pushed) — a code change belongs to
|
|
114
|
+
one. After you complete and go idle, treat `$LEGION_WORKSPACE` as read-only in every later
|
|
115
|
+
turn, including one an Envoy question starts: you are kept alive to answer questions, not to keep
|
|
116
|
+
editing. Do not create new commits, run `jj -R "$LEGION_WORKSPACE" new`, or touch tracked files
|
|
117
|
+
once your own handoff is committed (and, for the implementer, pushed) — a code change belongs to
|
|
118
|
+
whichever phase is active now. The same holds once the Go daemon has taken your phase back without
|
|
119
|
+
your completion: CI settling red while you tested, or reviewed a round you had not completed,
|
|
120
|
+
moves the issue to `implementing` and interrupts your turn, and the implementer starts only once
|
|
121
|
+
that turn has ended. Do not resume the interrupted work afterward.
|
|
118
122
|
|
|
119
123
|
On every start, and especially after revival or re-creation, read the issue and then the
|
|
120
124
|
committed predecessor handoffs in lifecycle order from
|
|
@@ -463,13 +467,18 @@ do:
|
|
|
463
467
|
Quote the answer verbatim in what you tell the architect: with the run and phase it names, the
|
|
464
468
|
difference between "my work is lost" and "my work belongs to the previous run" is visible.
|
|
465
469
|
|
|
466
|
-
**Stay in this session afterward.** Your process does not exit
|
|
467
|
-
|
|
468
|
-
|
|
469
|
-
|
|
470
|
-
|
|
471
|
-
|
|
472
|
-
|
|
470
|
+
**Stay in this session afterward.** Your process does not exit when your phase completes; it
|
|
471
|
+
stays live until your issue leaves the workflow. No move between phases stops it, and only an
|
|
472
|
+
explicit stop does: the issue closing; a move to `backlog`, `icebox` or `triage` (a child by a
|
|
473
|
+
person or its architect's `park_child`, or the root, which stops the whole tree); a child set back
|
|
474
|
+
to `todo` (which stops the interrupted phase's worker, not earlier roles); or an operator stop.
|
|
475
|
+
A stop keeps your session, and a crash relaunches it. Your next assignment arrives in this same
|
|
476
|
+
session. Other roles on
|
|
477
|
+
this issue may reach you through Envoy with
|
|
478
|
+
questions about the work you did — answer them, reading `$LEGION_WORKSPACE` and your own
|
|
479
|
+
committed handoff as needed, without mutating anything (see Workspace and handoff
|
|
480
|
+
precedence above). You will also be the one resumed, with a new prompt in this same
|
|
481
|
+
session, if this phase's work needs to run again.
|
|
473
482
|
|
|
474
483
|
When blocked on a product, scope, design, lifecycle, or cross-phase decision, `envoy_publish` the
|
|
475
484
|
owning architect a concise message: issue, phase, verified observation, what you tried, and the
|