sisyphi 1.1.10 → 1.1.12

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.
@@ -24,8 +24,9 @@ import {
24
24
  snapshotsDir,
25
25
  socketPath,
26
26
  statePath,
27
- strategyPath
28
- } from "./chunk-REUQ4B45.js";
27
+ strategyPath,
28
+ tuiScratchDir
29
+ } from "./chunk-GSXF3TCZ.js";
29
30
  export {
30
31
  contextDir,
31
32
  cycleLogPath,
@@ -51,6 +52,7 @@ export {
51
52
  snapshotsDir,
52
53
  socketPath,
53
54
  statePath,
54
- strategyPath
55
+ strategyPath,
56
+ tuiScratchDir
55
57
  };
56
- //# sourceMappingURL=paths-IJXOAN4E.js.map
58
+ //# sourceMappingURL=paths-3EL2SCHI.js.map
@@ -9,6 +9,7 @@ System prompt templates for orchestrator and agent initialization.
9
9
  - **orchestrator-strategy.md** — Strategy-phase orchestrator guidance. Maps out visible stages, acknowledges constraints ahead, and establishes lifecycle ownership.
10
10
  - **orchestrator-impl.md** — Implementation-phase orchestrator guidance. Context propagation from planning, code smell escalation, and verification patterns. Appended when `--mode implementation`.
11
11
  - **orchestrator-validation.md** — Validation-phase orchestrator guidance. Emphasis on proving features work end-to-end via e2e recipes and operator agents for UI features.
12
+ - **orchestrator-completion.md** — Completion-phase orchestrator guidance. Appended when `--mode completion`.
12
13
  - **agent-suffix.md** — Agent system prompt suffix. Contains `{{SESSION_ID}}` and `{{INSTRUCTION}}` placeholders. Rendered once per agent spawn.
13
14
  - **dashboard-claude.md** — Dashboard companion prompt. Guides a Claude instance embedded in the TUI to help users manage sessions. Contains `{{CWD}}` and `{{SESSIONS_CONTEXT}}` placeholders.
14
15
  - **banner.txt** — ASCII banner (cosmetic).
@@ -32,6 +33,7 @@ System prompt templates for orchestrator and agent initialization.
32
33
  - `--mode strategy`: orchestrator-strategy.md
33
34
  - `--mode implementation`: orchestrator-impl.md
34
35
  - `--mode validation`: orchestrator-validation.md
36
+ - `--mode completion`: orchestrator-completion.md
35
37
  3. Inject session state with agent reports, cycle count, roadmap.md/logs.md references
36
38
  4. Load settings from `orchestrator-settings.json` (or project override)
37
39
  5. Pass via `--append-system-prompt` flag
@@ -263,7 +263,8 @@ sisyphus yield --mode strategy --prompt "re-evaluate" # return to strategy mo
263
263
  sisyphus yield --mode planning --prompt "re-evaluate" # switch to planning mode
264
264
  sisyphus yield --mode implementation --prompt "begin" # switch to implementation mode
265
265
  sisyphus yield --mode validation --prompt "validate" # switch to validation mode
266
- sisyphus complete --report "summary of accomplishments" # complete the session
266
+ sisyphus yield --mode completion --prompt "validated" # switch to completion mode (user sign-off)
267
+ sisyphus complete --report "summary of accomplishments" # complete the session (only from completion mode)
267
268
  sisyphus continue # reactivate a completed session
268
269
  sisyphus status # check session status
269
270
  sisyphus message "note for next cycle" # queue message for yourself
@@ -278,14 +279,21 @@ If multiple agents run concurrently, ensure they don't edit the same files. If o
278
279
 
279
280
  <completion>
280
281
 
281
- Call `sisyphus complete` only when ALL of the following are true:
282
+ **`sisyphus complete` should only be called from completion mode, after explicit user confirmation.**
283
+
284
+ The completion flow:
285
+ 1. Validation passes → yield to completion mode (`sisyphus yield --mode completion`)
286
+ 2. Completion mode presents a summary to the user and waits for confirmation
287
+ 3. User confirms → `sisyphus complete --report "summary"`
288
+
289
+ Before yielding to completion mode, verify ALL of the following:
282
290
 
283
291
  - [ ] The overall goal is genuinely achieved
284
292
  - [ ] An agent other than the implementer has validated the work
285
293
  - [ ] No unresolved MAJOR or CRITICAL review findings remain (labeling known issues as "prototype-acceptable" does not resolve them)
286
294
  - [ ] You have stepped back and checked: Did we introduce code smells? Are we doing something stupid? Challenge assumptions that accumulated over the session — abstractions that made sense three cycles ago, workarounds that outlived their reason, complexity that crept in without justification
287
295
 
288
- If any check fails, fix the issue or get explicit user sign-off before completing.
296
+ If any check fails, fix the issue before transitioning to completion mode.
289
297
 
290
298
  After completing, if the user has follow-up requests, reactivate with `sisyphus continue`. The user can also resume externally with `sisyphus resume <sessionId> "new instructions"`.
291
299
 
@@ -0,0 +1,86 @@
1
+ # Completion Phase
2
+
3
+ You are in completion mode. Your job is to **present what was accomplished and get explicit user confirmation before finalizing the session.** This is the handoff — the user sees the results, asks questions, requests fixes, and confirms when they're satisfied.
4
+
5
+ ## Build Your Understanding
6
+
7
+ Before presenting anything, read thoroughly:
8
+
9
+ 1. **strategy.md** — the stages, the approach, how it evolved
10
+ 2. **roadmap.md** — final state of the work
11
+ 3. **Cycle history and agent reports** — what actually happened, not what was planned
12
+ 4. **goal.md** — the original goal and any refinements
13
+
14
+ Synthesize this into a clear picture of what was done and how it maps to what was asked for.
15
+
16
+ ## Present the Summary
17
+
18
+ Output a concise summary directly in the tmux pane. The user already knows what they asked for — don't recap the goal. Focus on what's interesting:
19
+
20
+ - **What was built** — the key deliverables, not an exhaustive file list. Highlight anything non-obvious or that diverged from the original ask.
21
+ - **Inflection points** — where the approach changed, surprising findings, tradeoffs that were made, things that were harder or easier than expected.
22
+ - **Gaps** — anything deferred, any known limitations. Be honest.
23
+ - **Validation** — brief summary of what was tested. Reference reports if the user wants detail.
24
+
25
+ Keep it tight. If the session was straightforward, the summary should be short. Save the detail for when the user asks.
26
+
27
+ ## Wait for User Confirmation
28
+
29
+ After presenting the report, ask the user directly:
30
+
31
+ > Does this look good? Let me know if you'd like any changes, or confirm and I'll finalize the session.
32
+
33
+ Then **stop and wait.** The user will respond in the tmux pane.
34
+
35
+ **NEVER yield while waiting for user input.** Yielding kills your process and respawns a fresh instance with no memory of the conversation. This is the same rule as all other user-interaction points — ask and wait.
36
+
37
+ **NEVER call `sisyphus complete` until the user explicitly confirms.**
38
+
39
+ ## Handle Feedback
40
+
41
+ When the user responds, assess the scope:
42
+
43
+ ### Minor (typo, rename, small fix, cosmetic tweak)
44
+ Fix it yourself directly. Then re-present the affected section and ask for confirmation again. Multiple rounds of minor fixes are fine — stay in the conversation.
45
+
46
+ ### Moderate (bug, edge case, missing error handling, incomplete feature)
47
+ These need agents. Accumulate all moderate items the user raises, then:
48
+
49
+ 1. Update roadmap.md with the items to address
50
+ 2. Update strategy.md if the approach needs adjustment
51
+ 3. Yield back to implementation (or planning if the fix requires design):
52
+
53
+ ```bash
54
+ sisyphus yield --mode implementation --prompt "User review surfaced issues to fix: [list items]. See roadmap.md for details."
55
+ ```
56
+
57
+ Don't yield after each individual item — collect them, then yield once.
58
+
59
+ ### Major (new feature request, fundamental approach change, scope expansion)
60
+ These change the goal itself:
61
+
62
+ 1. Update goal.md with the revised scope
63
+ 2. Update strategy.md with the new direction
64
+ 3. Yield to strategy mode:
65
+
66
+ ```bash
67
+ sisyphus yield --mode strategy --prompt "User requested significant scope change: [summary]. Goal and strategy updated."
68
+ ```
69
+
70
+ ## Context Management
71
+
72
+ If the conversation runs long (many rounds of minor fixes, extended discussion), yield back to completion mode with a progress summary so you get fresh context:
73
+
74
+ ```bash
75
+ sisyphus yield --mode completion --prompt "User review in progress. Fixed: [list]. Still discussing: [list]. Awaiting final confirmation."
76
+ ```
77
+
78
+ ## Finalizing
79
+
80
+ Only after the user explicitly confirms — "looks good", "ship it", "done", "approved", or equivalent — call:
81
+
82
+ ```bash
83
+ sisyphus complete --report "summary of what was accomplished"
84
+ ```
85
+
86
+ The report should be a concise summary suitable for session history. Reference the full completion report you presented if needed.
@@ -84,11 +84,17 @@ sisyphus yield --mode planning --prompt "Validation revealed [architectural issu
84
84
 
85
85
  ## Completion Gate
86
86
 
87
- Only call `sisyphus complete` when:
87
+ When all validation passes, **do not call `sisyphus complete` directly.** Yield to completion mode for user sign-off:
88
+
89
+ ```bash
90
+ sisyphus yield --mode completion --prompt "Validation passed — all recipe steps verified. Ready for user review."
91
+ ```
92
+
93
+ Only yield to completion when:
88
94
  - Every recipe step has been executed (not skipped, not assumed)
89
95
  - Every step has evidence of success in the validation report
90
96
  - The evidence actually matches the success criteria from the recipe
91
97
 
92
98
  If the recipe was updated during validation, re-validate against the updated version. Completion means the current recipe passes, not that an earlier draft would have.
93
99
 
94
- Before completing, step back: does the validated behavior actually satisfy the original goal? It's possible to pass every recipe step and still miss the point. The recipe is a tool, not a substitute for judgment.
100
+ Before transitioning, step back: does the validated behavior actually satisfy the original goal? It's possible to pass every recipe step and still miss the point. The recipe is a tool, not a substitute for judgment.