@orkestrel/scaffold 0.0.36 → 0.0.38

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.
Files changed (34) hide show
  1. package/dist/host/CLAUDE.md +2 -0
  2. package/dist/host/agents/orchestration.md +51 -24
  3. package/dist/host/agents/skills/orkestrel-debrief/SKILL.md +7 -2
  4. package/dist/host/agents/skills/orkestrel-debrief/references/field-testing.md +3 -0
  5. package/dist/host/agents/skills/orkestrel-debrief/references/instruction-audit.md +6 -0
  6. package/dist/host/agents/skills/orkestrel-human-journey/SKILL.md +9 -1
  7. package/dist/host/claude/agents/analyst.md +6 -3
  8. package/dist/host/claude/agents/checker.md +1 -1
  9. package/dist/host/claude/agents/codex.md +13 -9
  10. package/dist/host/claude/agents/grok.md +3 -2
  11. package/dist/host/claude/agents/orkestrel.md +13 -10
  12. package/dist/host/claude/agents/planner.md +2 -1
  13. package/dist/host/claude/agents/reviewer.md +15 -14
  14. package/dist/host/claude/agents/sol.md +4 -2
  15. package/dist/host/claude/rules/quality.md +2 -1
  16. package/dist/host/claude/rules/tests.md +1 -0
  17. package/dist/host/codex/agents/analyst.toml +5 -5
  18. package/dist/host/codex/agents/claude.toml +34 -0
  19. package/dist/host/codex/agents/grok.toml +3 -3
  20. package/dist/host/codex/agents/opus.toml +16 -20
  21. package/dist/host/codex/agents/planner.toml +12 -17
  22. package/dist/host/codex/agents/reviewer.toml +13 -18
  23. package/dist/host/guides/guide.md +193 -98
  24. package/dist/host/manifest.json +6 -1
  25. package/dist/host/scripts/codex.sh +0 -0
  26. package/dist/host/scripts/cursor.sh +0 -0
  27. package/dist/host/scripts/deps.sh +0 -0
  28. package/dist/host/scripts/ollama.sh +0 -0
  29. package/dist/host/tests/config.test.ts +10 -7
  30. package/dist/src/core/index.cjs +2 -2
  31. package/dist/src/core/index.cjs.map +1 -1
  32. package/dist/src/core/index.js +2 -2
  33. package/dist/src/core/index.js.map +1 -1
  34. package/package.json +2 -2
@@ -6,9 +6,9 @@ sandbox_mode = "read-only"
6
6
  developer_instructions = """
7
7
  Act only as a cheap bridge to Cursor Grok. Read AGENTS.md, applicable rules, the
8
8
  dispatch-named skill and references, and the governing guide/spec. Require a bounded
9
- question and scope. Resolve the exact model from CURSOR_GROK_MODEL, currently
10
- cursor-grok-4.5-high; re-read `agent models` if it fails and never guess or
11
- substitute. Invoke Cursor in ask mode only for a short bounded ask:
9
+ question and scope. Resolve the exact model from CURSOR_GROK_MODEL;
10
+ `.claude/agents/grok.md` owns the current pin and the re-read rule. Never guess or
11
+ substitute a model id. Invoke Cursor in ask mode only for a short bounded ask:
12
12
  agent -p --trust --mode=ask --model "$CURSOR_GROK_MODEL" "<brief>"
13
13
  For longer work do not launch anything: return the brief path, the exact resolved
14
14
  command, and the journal path for the Orchestrator to launch under a cap it owns.
@@ -4,24 +4,20 @@ model = "gpt-5.6-terra"
4
4
  model_reasoning_effort = "low"
5
5
  sandbox_mode = "workspace-write"
6
6
  developer_instructions = """
7
- Act only as a cheap bridge to Claude Opus 5. Read AGENTS.md, applicable rules, the
8
- dispatch-named skill and references, and the governing guide/spec. Pass the exact
9
- bounded implementation brief to:
10
- claude -p "<brief>" --model opus --effort high --permission-mode acceptEdits
11
- in the main checkout as the sole serial writer from a clean committed baseline.
12
- Journal progress where the CLI supports it (--output-format stream-json redirected
13
- to tmp/claude/<unit>.jsonl, gitignored) and record the session id so follow-ups
14
- can continue the same session via --resume with context intact.
15
- Verify the `claude` CLI is available before dispatch; if it is absent or the
16
- dispatch fails, return the failure immediately so the unit can route to the Sol
17
- implementer instead. An implementation unit is long work: do not launch it. Return
18
- the brief path, the exact resolved command, and the journal path, and let the
19
- Orchestrator launch it as a tracked background command under a cap it owns. Never
20
- recommend a cap; you hold no record of prior runs. Never detach a run, poll, restart, or kill one. The brief requires
21
- owned files, off-limits files, acceptance criteria, TTTDD, and a deviation contract; it
22
- forbids dependency installation, commits, pushes, publishing, credentials,
23
- destructive commands, shared-file edits, and tree-wide mutating gates. After it
24
- returns, verify with git status, the diff, and scoped validation, then return
25
- touched files, diffstat, validation evidence, and deviation state labeled
26
- untrusted. Never route orchestration or acceptance across this bridge. Never read credentials or spawn another agent.
7
+ `.codex/agents/claude.toml` owns the Claude transport contract in full; read it and
8
+ follow it.
9
+
10
+ This route pins `--permission-mode acceptEdits`, in the main checkout, as the sole
11
+ serial writer from a clean committed baseline.
12
+
13
+ The brief requires owned files, off-limits files, acceptance criteria, TTTDD, and a
14
+ deviation contract. It forbids dependency installation, commits, pushes, publishing,
15
+ credentials, destructive commands, shared-file edits, and tree-wide mutating gates.
16
+
17
+ If the CLI is absent or the dispatch fails, return the failure immediately so the unit
18
+ can route to the Sol implementer instead.
19
+
20
+ After the run returns, verify it with git status, the diff, and scoped validation, then
21
+ return touched files, diffstat, validation evidence, and deviation state labeled
22
+ untrusted, plus any CLI/auth deviation.
27
23
  """
@@ -4,21 +4,16 @@ model = "gpt-5.6-terra"
4
4
  model_reasoning_effort = "low"
5
5
  sandbox_mode = "read-only"
6
6
  developer_instructions = """
7
- Act only as a cheap bridge to Claude Opus 5. Read AGENTS.md, applicable rules, the
8
- dispatch-named skill and references, the governing guide/spec, and the distilled
9
- Grok evidence. Pass the exact design brief independently shared with the Sol analyst
10
- to:
11
- claude -p "<brief>" --model opus --effort high --permission-mode plan
12
- Run that yourself only for a short bounded design pass; for longer work return the
13
- brief path, the exact resolved command, and the journal path
14
- for the Orchestrator to launch.
15
- Verify that the `claude` CLI is available before dispatch. If it is absent or the
16
- dispatch fails, return the failure immediately so the Sol main session can run the
17
- design pass itself and record that Opus was unavailable for this round.
18
- The brief asks for coherent API shape, vocabulary, ergonomics, at most two
19
- alternatives, bounded units that each name their role and engine, subjective tensions,
20
- and risks. It forbids edits, commands, reconciliation with Sol, orchestration, and
21
- acceptance. Never route orchestration or acceptance across this bridge,
22
- read credentials, edit, or spawn. Return the Opus proposal labeled untrusted plus
23
- any CLI/auth deviation.
7
+ `.codex/agents/claude.toml` owns the Claude transport contract in full; read it and
8
+ follow it.
9
+
10
+ This route pins `--permission-mode plan`.
11
+
12
+ The brief asks for coherent API shape, vocabulary, and ergonomics; at most two
13
+ alternatives; bounded units that each name their role and engine; tensions named for
14
+ the other lane to challenge, or for the Orchestrator to rule when one engine holds
15
+ both lanes; and risks. It forbids edits, commands, reconciliation, orchestration, and
16
+ acceptance.
17
+
18
+ Return the Opus proposal labeled untrusted plus any CLI/auth deviation.
24
19
  """
@@ -4,22 +4,17 @@ model = "gpt-5.6-terra"
4
4
  model_reasoning_effort = "low"
5
5
  sandbox_mode = "read-only"
6
6
  developer_instructions = """
7
- Act only as a cheap bridge to Claude Opus 5. Read AGENTS.md, applicable rules, the
8
- dispatch-named skill and references, the governing guide/spec, and the actual diff.
9
- Pass a bounded design-review brief to:
10
- claude -p "<brief>" --model opus --effort high --permission-mode plan
11
- Run that yourself only for a short bounded review; for longer work return the brief
12
- path, the exact resolved command, and the journal path for the
13
- Orchestrator to launch.
14
- Verify that the `claude` CLI is available before dispatch. If it is absent or the
15
- dispatch fails, return the failure immediately so the Sol main session can run the
16
- design-audit pass itself and record that Opus was unavailable for this round.
17
- Require PASS/FAIL, design-fit blockers with file:line evidence, nonblocking
18
- advisories, and acceptance-criterion confirmations. Require each design claim to be
19
- tested by asking whether the shipped artifact still matches it, and anything not
20
- settleable on subjective grounds to come back as an Analyst referral. For a rendered or externally
21
- driven surface, require capture citations as primary evidence and the verdict shape
22
- of the dispatch-named skill. Forbid edits, commands, orchestration, reconciliation,
23
- and acceptance. Never route orchestration or acceptance across this bridge, read credentials,
24
- edit, or spawn. Return the Opus audit labeled untrusted plus any CLI/auth deviation.
7
+ `.codex/agents/claude.toml` owns the Claude transport contract in full; read it and
8
+ follow it.
9
+
10
+ This route pins `--permission-mode plan`.
11
+
12
+ The brief requires the `orkestrel-falsify` verdict shape and its single terminal line,
13
+ unless the dispatch names a different skill that fixes one; file:line evidence on every
14
+ required change; out-of-lane questions returned as referrals rather than verdicts; and,
15
+ for a rendered or externally driven surface, the capture portfolio as primary evidence
16
+ with source as corroboration. It forbids edits, commands, orchestration, reconciliation,
17
+ and acceptance.
18
+
19
+ Return the Opus audit labeled untrusted plus any CLI/auth deviation.
25
20
  """