alphacouncil-agent 0.9.0-solo-test.3 → 0.9.1

package/.claude/commands/alpha.md

@@ -13,7 +13,7 @@ Request: **$ARGUMENTS**
 |---|---|
 | *(empty)* | Print the mode table below and stop. Do not start a run. |
 | a ticker alone, or a question | **Full council.** Go to "Full council" below. |
-| ticker + `quick` | The `quick` evidence preset: 4 analysts + the selected master methods + debate. No verification. |
+| ticker + `quick` | The plugin-managed headless `quick_v1` council: 4 fixed analysts, 1-4 selected methods, one parallel bull/bear round and a short PM inside a hard 10-minute ceiling. |
 | ticker + `screen` | `screen_ticker` only. No language-model judgment, no subagents. |
 | ticker + `options` | `get_options_chain` only. |
 | ticker + `news` | `get_news` on the symbol, and on its CIK when it is a US filer. |
@@ -23,7 +23,7 @@ When `$ARGUMENTS` is empty, print exactly this and stop:
 
 ```
 /alpha <TICKER>          full council — shows every master, then asks which to run
-/alpha <TICKER> quick    4 analysts + selected masters + debate, no verification
+/alpha <TICKER> quick    quick_v1: 4 analysts incl. news + 1-4 masters + 1 parallel debate round (<=10m)
 /alpha <TICKER> screen   mechanical filings screen only        (no model spend)
 /alpha <TICKER> options  IV term structure, skew, positioning  (no model spend)
 /alpha <TICKER> news     dated filings and headlines           (no model spend)
@@ -38,34 +38,40 @@ Any listed equity. Filings-based modes need a US filer; for other markets say wh
 is and use `market_coverage`, rather than returning nothing.
 
 Say plainly that the four marked modes call keyless data tools and spawn no subagents, so
-they cost nothing beyond this turn. The council modes spawn one subagent per seat, and that
-is where the spend is.
+they cost nothing beyond this turn. Council modes can launch model workers; selected method
+seats may instead complete deterministically or return a recorded `out_of_scope` result.
 
 ## Council selection gate — mandatory for full and quick
 
 This gate happens before any research tool, run envelope, network fetch or subagent. It is
 the same in Claude Code, Codex, OpenCode and Grok Build.
 
-1. Call `begin_council_selection` with the symbol, original request, language and host. If
+1. Call `begin_council_selection` with the symbol, original request, language, host and the
+   intended `council_mode` (`full` by default; exactly `quick` for quick mode). If
    the request explicitly names masters, also pass those stable IDs as
    `preselected_master_ids`; this highlights them but does not confirm them.
 2. Show **every returned master individually, in the returned order and with its stable
    number**. Each row must include `identity`, `method`, `best_for` and `maturity`; a school
    name or a count is not a substitute for the individual catalog.
-3. Ask for one submission. Accept one number from `1..N`, any comma/space-separated
-   combination, ranges such as `1-4` or `1..4`, stable IDs/names, or `all`. A host-native
-   multi-select is a convenience only. If it is unavailable or cannot show the full catalog,
-   use the numbered text table and plain reply on every host.
-4. If the original request already named masters or said `all`, prefill that choice but
+3. Ask for one submission. In full mode accept one number from `1..N`, combinations,
+   ranges, stable IDs/names, or `all`. In quick mode the same complete 26-seat catalog is
+   displayed, but the submission must contain **1..4 seats** and `all` / `select_all` is
+   forbidden. A host-native multi-select is a convenience only. If it is unavailable or
+   cannot show the full catalog, use the numbered text table and plain reply on every host.
+4. If the original full-mode request already named masters or said `all`, prefill that choice but
    **still show the full catalog and require this run's submission**. Do not silently reuse a
-   prior choice. The submitted choice is the confirmation; do not ask a second confirmation.
+   prior choice. For quick, prefill at most four named methods and ask the user to reduce an
+   oversized/`all` request to 1-4. The submitted choice is the confirmation; do not ask a
+   second confirmation.
 5. Call `confirm_master_selection` with the returned `selection_id`, `catalog_hash`,
    `display_ack: true`, and exactly one of `selected_master_ids`, `select_all: true`, or
    `selection`. Retain the returned one-use `selection_receipt`.
-6. Only now call `plan_visible_run`, `collect_evidence` or `analyze_symbol`, passing that
-   `selection_receipt`. Do not also pass `masters` or `masters_roster`; the receipt is the
-   authoritative selection. If the receipt is missing, expired, stale or consumed, restart
-   from `begin_council_selection`.
+6. Only now call `plan_visible_run` (full only), `collect_evidence` (full only), or
+   `analyze_symbol`, passing that `selection_receipt` and the same symbol, prompt, language
+   and `council_mode`. Quick must use `analyze_symbol`. Do not also pass `masters` or
+   `masters_roster`; the receipt is authoritative and mode-bound. A full receipt cannot start
+   quick and a quick receipt cannot start full. For missing, expired, stale or consumed
+   receipts, or for a mode mismatch, restart from `begin_council_selection`.
 
 `screen`, `options`, `news` and `market` are data-only modes and skip this gate. `quick` is
 still a council judgment, so it never skips the gate; it changes the analyst fan-out, not the
@@ -85,13 +91,42 @@ master-selection contract.
    conclusion, not an abstention.
 5. Give price bands with the condition attached to each. "The cycle position is undetermined"
    changes what the bands depend on; it does not excuse leaving them out.
-
-## Quick mode owes the reader one sentence
-
-Complete the same mandatory council selection gate, then use four analysts, the selected
-masters and the debate without verifier fan-out. State that evidence coverage and
-cross-verification are thinner than a standard run; list the selected masters so the quick
-result cannot be mistaken for the full bench.
+6. Full mode fails closed at the evidence barrier. If any mandatory evidence role still
+   fails after its bounded parse repair, do not spend more model calls on masters, bull/bear
+   or PM. Persist an `incomplete` run naming the failed evidence and skipped downstream roles.
+
+## Quick mode — plugin-managed headless quick_v1
+
+Quick is an explicit, bounded council contract, not an automatic downgrade of full and not a
+visible-host orchestration. `plan_visible_run` rejects `council_mode=quick`; run it through
+plugin-managed headless `analyze_symbol` so the server can enforce the deadline.
+
+1. Complete Stage 0 with `council_mode: "quick"`; display all 26 methods and confirm 1-4.
+2. Call `analyze_symbol` once with the exact same prompt/language/mode and one-use receipt,
+   `dry_run:false`, `wait_for_completion:false`, and optionally `total_timeout_ms` no greater
+   than `600000`. Do not pass `tasks`: quick fixes and starts these four roles in parallel:
+   `market_data`, `earnings_deep_dive`, `valuation_long_short`,
+   `news_industry_management`.
+3. The news role covers the highest-impact company and industry developments from the 120
+   days ending at `as_of`; future, undated and older items are excluded from recent news.
+4. The hard end-to-end budget starts with the durable queued run and includes a 20-second
+   grounding wait, evidence workers capped at 210 seconds each, up to four master workers in
+   parallel capped at 90 seconds each, one bull/bear statement round in parallel capped at
+   90 seconds per side, one short PM capped at 90 seconds, retries and a 20-second
+   finalization reserve. Callers may lower the total but never raise it above 600000 ms.
+5. No round-2 rebuttal, round-3 exact Q&A or adversarial verifier fan-out runs. Scoped source
+   IDs are still checked. The report uses `quick_v1`, not `full_v2`, and
+   `full_council_equivalent` is false.
+6. Poll the same run with `read_run`; never create a replacement run. Terminal statuses are
+   `complete`, `degraded`, `incomplete`, `needs_verification`, `needs_revision`, and `failed`.
+7. `degraded` is a real quick terminal state, not success renamed: at least 2 of 4 evidence
+   roles completed, every failed role has a sanitized recorded packet/diagnostic, at least
+   one bull/bear side completed, the PM completed and all selected methods were recorded.
+   The system-owned degraded ledger and handoff must name every degraded task/side and cause.
+8. The concise handoff must include rating/winner/confidence, every selected stable master ID
+   and stance, all four analyst statuses and summaries, dated recent news, valuation/position,
+   data gaps and artifact paths. Method-seat output is a recorded lens result, never a quote
+   from the named person. State explicitly that quick is not equivalent to full council.
 
 ## Screen mode
 

package/.claude-plugin/marketplace.json

@@ -6,14 +6,14 @@
   },
   "metadata": {
     "description": "AlphaCouncil plugins for public-equity research.",
-    "version": "0.9.0-solo-test.3"
+    "version": "0.9.1"
   },
   "plugins": [
     {
       "name": "alphacouncil-agent",
       "source": "./",
       "description": "Multi-agent equity research: evidence packets, bull/bear debate, portfolio-manager decision.",
-      "version": "0.9.0-solo-test.3",
+      "version": "0.9.1",
       "author": {
         "name": "Zhao73"
       },

package/.claude-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "alphacouncil-agent",
-  "version": "0.9.0-solo-test.3",
+  "version": "0.9.1",
   "description": "Multi-agent public-equity research workflow for Claude Code: spawns analyst workers, gathers sourced JSON evidence packets, runs bull/bear debate, and produces a portfolio-manager Buy/Overweight/Hold/Underweight/Sell decision with a full report.",
   "author": {
     "name": "Zhao73"

package/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "alphacouncil-agent",
-  "version": "0.9.0-solo-test.3",
+  "version": "0.9.1",
   "description": "Host-visible or background public-equity research workflow with shared evidence packets.",
   "author": {
     "name": "Zhao73"

package/.grok/commands/alpha.md

@@ -13,7 +13,7 @@ Request: **$ARGUMENTS**
 |---|---|
 | *(empty)* | Print the mode table below and stop. Do not start a run. |
 | a ticker alone, or a question | **Full council.** Go to "Full council" below. |
-| ticker + `quick` | The `quick` evidence preset: 4 analysts + the selected master methods + debate. No verification. |
+| ticker + `quick` | The plugin-managed headless `quick_v1` council: 4 fixed analysts, 1-4 selected methods, one parallel bull/bear round and a short PM inside a hard 10-minute ceiling. |
 | ticker + `screen` | `screen_ticker` only. No language-model judgment, no subagents. |
 | ticker + `options` | `get_options_chain` only. |
 | ticker + `news` | `get_news` on the symbol, and on its CIK when it is a US filer. |
@@ -23,7 +23,7 @@ When `$ARGUMENTS` is empty, print exactly this and stop:
 
 ```
 /alpha <TICKER>          full council — shows every master, then asks which to run
-/alpha <TICKER> quick    4 analysts + selected masters + debate, no verification
+/alpha <TICKER> quick    quick_v1: 4 analysts incl. news + 1-4 masters + 1 parallel debate round (<=10m)
 /alpha <TICKER> screen   mechanical filings screen only        (no model spend)
 /alpha <TICKER> options  IV term structure, skew, positioning  (no model spend)
 /alpha <TICKER> news     dated filings and headlines           (no model spend)
@@ -38,34 +38,40 @@ Any listed equity. Filings-based modes need a US filer; for other markets say wh
 is and use `market_coverage`, rather than returning nothing.
 
 Say plainly that the four marked modes call keyless data tools and spawn no subagents, so
-they cost nothing beyond this turn. The council modes spawn one subagent per seat, and that
-is where the spend is.
+they cost nothing beyond this turn. Council modes can launch model workers; selected method
+seats may instead complete deterministically or return a recorded `out_of_scope` result.
 
 ## Council selection gate — mandatory for full and quick
 
 This gate happens before any research tool, run envelope, network fetch or subagent. It is
 the same in Claude Code, Codex, OpenCode and Grok Build.
 
-1. Call `begin_council_selection` with the symbol, original request, language and host. If
+1. Call `begin_council_selection` with the symbol, original request, language, host and the
+   intended `council_mode` (`full` by default; exactly `quick` for quick mode). If
    the request explicitly names masters, also pass those stable IDs as
    `preselected_master_ids`; this highlights them but does not confirm them.
 2. Show **every returned master individually, in the returned order and with its stable
    number**. Each row must include `identity`, `method`, `best_for` and `maturity`; a school
    name or a count is not a substitute for the individual catalog.
-3. Ask for one submission. Accept one number from `1..N`, any comma/space-separated
-   combination, ranges such as `1-4` or `1..4`, stable IDs/names, or `all`. A host-native
-   multi-select is a convenience only. If it is unavailable or cannot show the full catalog,
-   use the numbered text table and plain reply on every host.
-4. If the original request already named masters or said `all`, prefill that choice but
+3. Ask for one submission. In full mode accept one number from `1..N`, combinations,
+   ranges, stable IDs/names, or `all`. In quick mode the same complete 26-seat catalog is
+   displayed, but the submission must contain **1..4 seats** and `all` / `select_all` is
+   forbidden. A host-native multi-select is a convenience only. If it is unavailable or
+   cannot show the full catalog, use the numbered text table and plain reply on every host.
+4. If the original full-mode request already named masters or said `all`, prefill that choice but
    **still show the full catalog and require this run's submission**. Do not silently reuse a
-   prior choice. The submitted choice is the confirmation; do not ask a second confirmation.
+   prior choice. For quick, prefill at most four named methods and ask the user to reduce an
+   oversized/`all` request to 1-4. The submitted choice is the confirmation; do not ask a
+   second confirmation.
 5. Call `confirm_master_selection` with the returned `selection_id`, `catalog_hash`,
    `display_ack: true`, and exactly one of `selected_master_ids`, `select_all: true`, or
    `selection`. Retain the returned one-use `selection_receipt`.
-6. Only now call `plan_visible_run`, `collect_evidence` or `analyze_symbol`, passing that
-   `selection_receipt`. Do not also pass `masters` or `masters_roster`; the receipt is the
-   authoritative selection. If the receipt is missing, expired, stale or consumed, restart
-   from `begin_council_selection`.
+6. Only now call `plan_visible_run` (full only), `collect_evidence` (full only), or
+   `analyze_symbol`, passing that `selection_receipt` and the same symbol, prompt, language
+   and `council_mode`. Quick must use `analyze_symbol`. Do not also pass `masters` or
+   `masters_roster`; the receipt is authoritative and mode-bound. A full receipt cannot start
+   quick and a quick receipt cannot start full. For missing, expired, stale or consumed
+   receipts, or for a mode mismatch, restart from `begin_council_selection`.
 
 `screen`, `options`, `news` and `market` are data-only modes and skip this gate. `quick` is
 still a council judgment, so it never skips the gate; it changes the analyst fan-out, not the
@@ -85,13 +91,42 @@ master-selection contract.
    conclusion, not an abstention.
 5. Give price bands with the condition attached to each. "The cycle position is undetermined"
    changes what the bands depend on; it does not excuse leaving them out.
-
-## Quick mode owes the reader one sentence
-
-Complete the same mandatory council selection gate, then use four analysts, the selected
-masters and the debate without verifier fan-out. State that evidence coverage and
-cross-verification are thinner than a standard run; list the selected masters so the quick
-result cannot be mistaken for the full bench.
+6. Full mode fails closed at the evidence barrier. If any mandatory evidence role still
+   fails after its bounded parse repair, do not spend more model calls on masters, bull/bear
+   or PM. Persist an `incomplete` run naming the failed evidence and skipped downstream roles.
+
+## Quick mode — plugin-managed headless quick_v1
+
+Quick is an explicit, bounded council contract, not an automatic downgrade of full and not a
+visible-host orchestration. `plan_visible_run` rejects `council_mode=quick`; run it through
+plugin-managed headless `analyze_symbol` so the server can enforce the deadline.
+
+1. Complete Stage 0 with `council_mode: "quick"`; display all 26 methods and confirm 1-4.
+2. Call `analyze_symbol` once with the exact same prompt/language/mode and one-use receipt,
+   `dry_run:false`, `wait_for_completion:false`, and optionally `total_timeout_ms` no greater
+   than `600000`. Do not pass `tasks`: quick fixes and starts these four roles in parallel:
+   `market_data`, `earnings_deep_dive`, `valuation_long_short`,
+   `news_industry_management`.
+3. The news role covers the highest-impact company and industry developments from the 120
+   days ending at `as_of`; future, undated and older items are excluded from recent news.
+4. The hard end-to-end budget starts with the durable queued run and includes a 20-second
+   grounding wait, evidence workers capped at 210 seconds each, up to four master workers in
+   parallel capped at 90 seconds each, one bull/bear statement round in parallel capped at
+   90 seconds per side, one short PM capped at 90 seconds, retries and a 20-second
+   finalization reserve. Callers may lower the total but never raise it above 600000 ms.
+5. No round-2 rebuttal, round-3 exact Q&A or adversarial verifier fan-out runs. Scoped source
+   IDs are still checked. The report uses `quick_v1`, not `full_v2`, and
+   `full_council_equivalent` is false.
+6. Poll the same run with `read_run`; never create a replacement run. Terminal statuses are
+   `complete`, `degraded`, `incomplete`, `needs_verification`, `needs_revision`, and `failed`.
+7. `degraded` is a real quick terminal state, not success renamed: at least 2 of 4 evidence
+   roles completed, every failed role has a sanitized recorded packet/diagnostic, at least
+   one bull/bear side completed, the PM completed and all selected methods were recorded.
+   The system-owned degraded ledger and handoff must name every degraded task/side and cause.
+8. The concise handoff must include rating/winner/confidence, every selected stable master ID
+   and stance, all four analyst statuses and summaries, dated recent news, valuation/position,
+   data gaps and artifact paths. Method-seat output is a recorded lens result, never a quote
+   from the named person. State explicitly that quick is not equivalent to full council.
 
 ## Screen mode
 

package/.opencode/command/alpha.md

@@ -13,7 +13,7 @@ Request: **$ARGUMENTS**
 |---|---|
 | *(empty)* | Print the mode table below and stop. Do not start a run. |
 | a ticker alone, or a question | **Full council.** Go to "Full council" below. |
-| ticker + `quick` | The `quick` evidence preset: 4 analysts + the selected master methods + debate. No verification. |
+| ticker + `quick` | The plugin-managed headless `quick_v1` council: 4 fixed analysts, 1-4 selected methods, one parallel bull/bear round and a short PM inside a hard 10-minute ceiling. |
 | ticker + `screen` | `screen_ticker` only. No language-model judgment, no subagents. |
 | ticker + `options` | `get_options_chain` only. |
 | ticker + `news` | `get_news` on the symbol, and on its CIK when it is a US filer. |
@@ -23,7 +23,7 @@ When `$ARGUMENTS` is empty, print exactly this and stop:
 
 ```
 /alpha <TICKER>          full council — shows every master, then asks which to run
-/alpha <TICKER> quick    4 analysts + selected masters + debate, no verification
+/alpha <TICKER> quick    quick_v1: 4 analysts incl. news + 1-4 masters + 1 parallel debate round (<=10m)
 /alpha <TICKER> screen   mechanical filings screen only        (no model spend)
 /alpha <TICKER> options  IV term structure, skew, positioning  (no model spend)
 /alpha <TICKER> news     dated filings and headlines           (no model spend)
@@ -38,34 +38,40 @@ Any listed equity. Filings-based modes need a US filer; for other markets say wh
 is and use `market_coverage`, rather than returning nothing.
 
 Say plainly that the four marked modes call keyless data tools and spawn no subagents, so
-they cost nothing beyond this turn. The council modes spawn one subagent per seat, and that
-is where the spend is.
+they cost nothing beyond this turn. Council modes can launch model workers; selected method
+seats may instead complete deterministically or return a recorded `out_of_scope` result.
 
 ## Council selection gate — mandatory for full and quick
 
 This gate happens before any research tool, run envelope, network fetch or subagent. It is
 the same in Claude Code, Codex, OpenCode and Grok Build.
 
-1. Call `begin_council_selection` with the symbol, original request, language and host. If
+1. Call `begin_council_selection` with the symbol, original request, language, host and the
+   intended `council_mode` (`full` by default; exactly `quick` for quick mode). If
    the request explicitly names masters, also pass those stable IDs as
    `preselected_master_ids`; this highlights them but does not confirm them.
 2. Show **every returned master individually, in the returned order and with its stable
    number**. Each row must include `identity`, `method`, `best_for` and `maturity`; a school
    name or a count is not a substitute for the individual catalog.
-3. Ask for one submission. Accept one number from `1..N`, any comma/space-separated
-   combination, ranges such as `1-4` or `1..4`, stable IDs/names, or `all`. A host-native
-   multi-select is a convenience only. If it is unavailable or cannot show the full catalog,
-   use the numbered text table and plain reply on every host.
-4. If the original request already named masters or said `all`, prefill that choice but
+3. Ask for one submission. In full mode accept one number from `1..N`, combinations,
+   ranges, stable IDs/names, or `all`. In quick mode the same complete 26-seat catalog is
+   displayed, but the submission must contain **1..4 seats** and `all` / `select_all` is
+   forbidden. A host-native multi-select is a convenience only. If it is unavailable or
+   cannot show the full catalog, use the numbered text table and plain reply on every host.
+4. If the original full-mode request already named masters or said `all`, prefill that choice but
    **still show the full catalog and require this run's submission**. Do not silently reuse a
-   prior choice. The submitted choice is the confirmation; do not ask a second confirmation.
+   prior choice. For quick, prefill at most four named methods and ask the user to reduce an
+   oversized/`all` request to 1-4. The submitted choice is the confirmation; do not ask a
+   second confirmation.
 5. Call `confirm_master_selection` with the returned `selection_id`, `catalog_hash`,
    `display_ack: true`, and exactly one of `selected_master_ids`, `select_all: true`, or
    `selection`. Retain the returned one-use `selection_receipt`.
-6. Only now call `plan_visible_run`, `collect_evidence` or `analyze_symbol`, passing that
-   `selection_receipt`. Do not also pass `masters` or `masters_roster`; the receipt is the
-   authoritative selection. If the receipt is missing, expired, stale or consumed, restart
-   from `begin_council_selection`.
+6. Only now call `plan_visible_run` (full only), `collect_evidence` (full only), or
+   `analyze_symbol`, passing that `selection_receipt` and the same symbol, prompt, language
+   and `council_mode`. Quick must use `analyze_symbol`. Do not also pass `masters` or
+   `masters_roster`; the receipt is authoritative and mode-bound. A full receipt cannot start
+   quick and a quick receipt cannot start full. For missing, expired, stale or consumed
+   receipts, or for a mode mismatch, restart from `begin_council_selection`.
 
 `screen`, `options`, `news` and `market` are data-only modes and skip this gate. `quick` is
 still a council judgment, so it never skips the gate; it changes the analyst fan-out, not the
@@ -85,13 +91,42 @@ master-selection contract.
    conclusion, not an abstention.
 5. Give price bands with the condition attached to each. "The cycle position is undetermined"
    changes what the bands depend on; it does not excuse leaving them out.
-
-## Quick mode owes the reader one sentence
-
-Complete the same mandatory council selection gate, then use four analysts, the selected
-masters and the debate without verifier fan-out. State that evidence coverage and
-cross-verification are thinner than a standard run; list the selected masters so the quick
-result cannot be mistaken for the full bench.
+6. Full mode fails closed at the evidence barrier. If any mandatory evidence role still
+   fails after its bounded parse repair, do not spend more model calls on masters, bull/bear
+   or PM. Persist an `incomplete` run naming the failed evidence and skipped downstream roles.
+
+## Quick mode — plugin-managed headless quick_v1
+
+Quick is an explicit, bounded council contract, not an automatic downgrade of full and not a
+visible-host orchestration. `plan_visible_run` rejects `council_mode=quick`; run it through
+plugin-managed headless `analyze_symbol` so the server can enforce the deadline.
+
+1. Complete Stage 0 with `council_mode: "quick"`; display all 26 methods and confirm 1-4.
+2. Call `analyze_symbol` once with the exact same prompt/language/mode and one-use receipt,
+   `dry_run:false`, `wait_for_completion:false`, and optionally `total_timeout_ms` no greater
+   than `600000`. Do not pass `tasks`: quick fixes and starts these four roles in parallel:
+   `market_data`, `earnings_deep_dive`, `valuation_long_short`,
+   `news_industry_management`.
+3. The news role covers the highest-impact company and industry developments from the 120
+   days ending at `as_of`; future, undated and older items are excluded from recent news.
+4. The hard end-to-end budget starts with the durable queued run and includes a 20-second
+   grounding wait, evidence workers capped at 210 seconds each, up to four master workers in
+   parallel capped at 90 seconds each, one bull/bear statement round in parallel capped at
+   90 seconds per side, one short PM capped at 90 seconds, retries and a 20-second
+   finalization reserve. Callers may lower the total but never raise it above 600000 ms.
+5. No round-2 rebuttal, round-3 exact Q&A or adversarial verifier fan-out runs. Scoped source
+   IDs are still checked. The report uses `quick_v1`, not `full_v2`, and
+   `full_council_equivalent` is false.
+6. Poll the same run with `read_run`; never create a replacement run. Terminal statuses are
+   `complete`, `degraded`, `incomplete`, `needs_verification`, `needs_revision`, and `failed`.
+7. `degraded` is a real quick terminal state, not success renamed: at least 2 of 4 evidence
+   roles completed, every failed role has a sanitized recorded packet/diagnostic, at least
+   one bull/bear side completed, the PM completed and all selected methods were recorded.
+   The system-owned degraded ledger and handoff must name every degraded task/side and cause.
+8. The concise handoff must include rating/winner/confidence, every selected stable master ID
+   and stance, all four analyst statuses and summaries, dated recent news, valuation/position,
+   data gaps and artifact paths. Method-seat output is a recorded lens result, never a quote
+   from the named person. State explicitly that quick is not equivalent to full council.
 
 ## Screen mode
 

package/AGENTS.md

@@ -8,17 +8,55 @@ For anti-laziness governance, also follow `skills/agent-skills-governance/SKILL.
 
 Run `npm run check` after any code or prompt change.
 
+## Current Release Boundary
+
+Package/plugin version `0.9.1` is a non-GA preview published to npm's `next` dist-tag and
+marked as a GitHub prerelease. Its build channel remains `solo_test`: 26 physical provisional
+PersonaPack v3 packs, 26 `operator_lens` seats, 52 `provisional_derived_proxy` tools,
+0 operational seats and 0 validated `method_model` seats. Do not present this feature
+release, a passing quick report or a packaged smoke as PersonaPack production GA.
+
+## Council Modes
+
+`full` is the default and uses the `full_v2` contract. It requires all eight mandatory
+evidence roles, every selected method, the three-round Bull/Bear cross-exam and PM. If any
+mandatory evidence role still fails after its one bounded parse-only repair, persist the
+failure/diagnostic artifacts, skip method/debate/PM model calls and terminate `incomplete`
+before downstream synthesis. Never auto-downgrade that run to quick.
+
+`quick` is explicit and uses `quick_v1`. It can run only through plugin-managed headless
+`analyze_symbol`; `plan_visible_run` rejects it. Quick launches the fixed four evidence roles
+`market_data`, `earnings_deep_dive`, `valuation_long_short`, and
+`news_industry_management` in parallel; 1-4 selected methods in parallel; one Bull and one
+Bear statement in parallel; then one short PM. It has no rebuttal/Q&A rounds and no
+adversarial `source_fidelity`/`rederivation`/`refuter` fan-out.
+
+Quick company/industry news must be dated within the 120 days ending at `as_of`. Future,
+undated and older items are gaps, not recent news. Its hard queue-to-persistence ceiling is
+600000 ms: grounding wait 20s; each evidence worker 210s; each selected-method worker 90s;
+Bull and Bear 90s per side; PM 90s; final assembly/persistence reserve 20s. Retry time is
+inside the same caps. The ceiling may be lowered, never raised.
+
+Quick may terminate `degraded` only under its documented minimum-coverage rule and one
+system-owned idempotent degraded ledger. `report_quality=passed` validates `quick_v1`
+structure only; it does not turn degraded into complete or imply full-council equivalence.
+Method-seat output is a recorded provisional lens result, never a quotation from the named
+person.
+
 ## Hosts
 
 The MCP server is the load-bearing integration on every host: it reads `personas/` directly,
 so a host that ignores the generated agent files still gets correct prompts.
 
 Every host also follows the same mandatory master-selection protocol for a full or quick
-council. Call `begin_council_selection`, display every returned entry with its number,
-identity, method and `best_for`, collect a numbered/range/ID/`all` submission, then call
-`confirm_master_selection` with `display_ack: true`. Existing names in the request are only
-a prefill; the full catalog is still shown. Only the returned one-use `selection_receipt`
-may authorize `plan_visible_run`, `collect_evidence` or `analyze_symbol`. A host-native
+council. Call `begin_council_selection` with the intended `council_mode`, display all 26
+returned entries with number, identity, method and `best_for`, collect one submission, then
+call `confirm_master_selection` with `display_ack: true`. Full accepts numbers, ranges,
+stable IDs or `all`; quick accepts exactly 1-4 distinct methods and rejects `all` and
+`select_all`. Existing names in the request are only a prefill; the full catalog is still
+shown. Only the returned one-use, mode-bound `selection_receipt`, reused with the same
+symbol, prompt, language and mode, may authorize the applicable execution tool. A full
+receipt cannot launch quick and a quick receipt cannot launch full. A host-native
 multi-select is optional UI sugar; the numbered text fallback is mandatory on Claude Code,
 Codex, OpenCode and Grok Build. Data-only `screen`, `options`, `news` and `market` modes are
 the only `/alpha` routes that skip this gate.
@@ -89,7 +127,7 @@ rather than four in a menu of a hundred.
 | Invocation | What runs | Model spend |
 |---|---|---|
 | `/alpha <ticker>` | Shows every master, confirms `1..N`/ranges/`all`, then runs the full council | one subagent per selected seat |
-| `/alpha <ticker> quick` | Same mandatory master selection, then 4 analysts + selected masters + debate; no verification | varies with selection |
+| `/alpha <ticker> quick` | Shows all 26, confirms 1-4 (no `all`), then plugin-managed `quick_v1` (≤10m) | varies with selection |
 | `/alpha <ticker> screen` | Mechanical filings screen only | **none** |
 | `/alpha <ticker> options` | IV term structure, skew, positioning | **none** |
 | `/alpha <ticker> news` | Dated filings and headlines | **none** |
@@ -98,7 +136,9 @@ rather than four in a menu of a hundred.
 
 The four marked **none** call keyless data tools and spawn no subagents, so they cost
 nothing beyond the turn you type them in. Full and quick are council modes: both require a
-fresh selection receipt, and their cost varies with the selected master count.
+fresh mode-bound selection receipt. Quick must poll the single durable `run_id` returned by
+`analyze_symbol(wait_for_completion=false)` through `read_run`; never emulate quick with
+visible agents or create a replacement run when it is slow.
 
 | Host | Where it reads them |
 |---|---|

package/CHANGELOG.md

@@ -2,6 +2,58 @@
 
 Notable changes per release. Dates are UTC.
 
+## [0.9.1] — 2026-07-28
+
+This is a **non-GA quick-council preview** published on npm's `next` dist-tag and as a
+GitHub prerelease. The bare SemVer does not promote the PersonaPack assurance level: the
+build profile remains `solo_test`, `production_eligible=false` and
+`method_model_eligible=false`; all 26 named-investor packs remain provisional
+`operator_lens`, not validated simulations of those people.
+
+### Added
+
+- A first-class `council_mode=quick` with a non-overridable ten-minute end-to-end ceiling.
+  It runs four fixed evidence seats in parallel (`market_data`, `earnings_deep_dive`,
+  `valuation_long_short`, `news_industry_management`), one to four selected method seats,
+  one parallel Bull/Bear statement, and one short PM synthesis.
+- `quick_v1`, a separate 13-section report contract. It records
+  `full_council_equivalent=false`, one expected debate round and
+  `adversarial_verification=not_run`; a passing quick report is never represented as a
+  passing full council.
+- A mode-bound selection receipt. Quick still displays the complete 26-seat catalog, but
+  accepts at most four explicit seats and rejects `all`; a quick receipt cannot start full.
+- An explicit terminal `degraded` state. A quick run may deliver when at least two evidence
+  seats and at least one debate side succeeded and the PM plus all selected methods were
+  recorded. A system-owned ledger names every failed/degraded seat and reason even when the
+  report-structure check passes.
+- A 120-day, `as_of`-bounded recent-news handoff. Undated, future and stale sources are
+  excluded and counted as gaps instead of being mislabeled as recent news.
+
+### Fixed
+
+- Full analysis now fails fast before method, debate and PM model calls when mandatory
+  evidence remains missing. The older RKLB runs spent another 16–43 minutes after the
+  result was already guaranteed to be incomplete.
+- Cross-seat prompts no longer resend raw transcripts or full report Markdown. Quick
+  evidence keeps a bounded claim/source projection with referential integrity; full PM
+  context retains bounded three-round summaries and exact Q&A while dropping artifact-only
+  raw text.
+- Worker and synthesis budgets share one quick deadline: grounding wait 20 seconds,
+  evidence up to 210 seconds in one four-seat wave, up to four method seats in one 90-second
+  wave, Bull/Bear in one parallel 90-second wave, PM up to 90 seconds, and 20 seconds
+  reserved for deterministic finalization. These are ceilings, not a success guarantee.
+- Failure fallback no longer emits a synthetic `Hold`; it records
+  `NEEDS_MANAGER_REVIEW`, `decision_available=false` and no rating.
+- `degraded` is recognized as terminal during polling and startup recovery. Evidence
+  telemetry uses `evidence_degraded` rather than the misleading `evidence_complete` event.
+- `npm publish --dry-run` no longer leaks `npm_config_dry_run` into the nested package-parity
+  `npm pack`, which previously returned metadata without creating the tarball and caused a
+  false release failure.
+
+The production GA gate remains intentionally failing and is not bypassed by this release.
+Formal GA requires a later monotonic version such as `0.10.0` or `1.0.0`, plus the existing
+source, method-model, experiment, physical-host and signed-release evidence.
+
 ## [0.9.0-solo-test.3] — 2026-07-27
 
 This is a post-acceptance hotfix for the **0.9.0 solo-test prerelease** channel. An