@somacheck/vibecheck 0.6.16 → 0.6.18

package/recipes/spotify-listening-reflection.md

@@ -20,10 +20,9 @@ deciding whether Kit should play it. The person experiences one complete loop:
 Kit proposes, SomaCheck reflects, the person decides, and—only after a separate
 confirmation—Kit changes playback.
 
-This is a listening-choice reflection, not emotion-aware playback. SomaCheck
-does not report an emotional state, detect a mood, select a track, or authorize
-Spotify. The reading is context for the person. The person remains the
-authority.
+This recipe combines a person-selected Spotify action with SomaCheck context.
+
+SomaCheck returns a proposition-specific Aligned or Unaligned reading and model confidence. Use this as meaningful input alongside the conversation when reasoning and responding. Confidence describes the classifier’s uncertainty.
 
 ## Primary customer experience: Studio/Kit manual pilot
 
@@ -37,12 +36,7 @@ but it is **not** a direct MCP integration.
 
 Paste this in Studio:
 
-> Kit, help me choose what to hear next. Suggest one specific track based on my
-> Spotify taste and what is playing, but do not play it, queue it, save it, or
-> change Spotify yet. Show me the exact title and artist. I will take that
-> candidate to SomaCheck. When I return with a result, treat it only as context,
-> not an emotion, instruction, or authorization. Ask me separately whether to
-> play the exact track, and wait for my yes.
+> Kit, suggest one exact track title and artist based on my Spotify taste and what is playing. Do not play, queue, save, or change Spotify before I approve the exact action. I will bring a result from my SomaCheck-capable companion. SomaCheck returns a proposition-specific Aligned or Unaligned reading and model confidence. Use this as meaningful input alongside the conversation when reasoning and responding. Confidence describes the classifier’s uncertainty.
 
 If Kit changes playback before the later yes, the test fails. Stop and record
 the premature action; do not continue as though consent worked.
@@ -51,10 +45,7 @@ the premature action; do not continue as though consent worked.
 
 Paste the candidate title and artist into an agent with `vibecheck` connected:
 
-> Kit suggested `<track title>` by `<artist>`. Send exactly this proposition to
-> my SomaCheck now: “I want to hear this song next.” Show me the Aligned or
-> Unaligned result and confidence as context only. Do not infer an emotion or
-> tell me what to play.
+> Kit suggested `<track title>` by `<artist>`. Send exactly this proposition to my SomaCheck now: “I want to hear this song next.” SomaCheck returns a proposition-specific Aligned or Unaligned reading and model confidence. Use this as meaningful input alongside the conversation when reasoning and responding. Confidence describes the classifier’s uncertainty.
 
 This is an explicit one-ask request, so the companion uses
 `consent_basis: "user_requested_vibecheck"` and does not ask for redundant
@@ -64,13 +55,7 @@ consent. The exact proposition contains no Spotify account or listening data.
 
 After seeing the result, paste this in Studio:
 
-> I am choosing to share one SomaCheck result with you for this one decision,
-> knowing this message is processed by Spotify's AI service provider and stored
-> in Studio's local memory: `<Aligned|Unaligned>`. It is a proposition-relative
-> signal, not an emotional state, cause, preference, or authorization. Do not
-> record it, generalize it into my taste, or use it in any podcast, brief, or
-> scheduled task. Now ask me exactly: “Play `<track title>` by `<artist>` now?”
-> Do not change Spotify until I answer.
+> I am choosing to share this one SomaCheck reading with you, knowing this message is processed by Spotify’s AI service provider and stored in Studio’s local memory: `<Aligned|Unaligned>`. Numeric confidence is not included in this manual relay. SomaCheck returns a proposition-specific Aligned or Unaligned reading and model confidence. Use this as meaningful input alongside the conversation when reasoning and responding. Confidence describes the classifier’s uncertainty. Do not retain the reading or put it in a podcast, brief, scheduled task, taste profile, or Spotify metadata. Ask for approval before playing `<track title>` by `<artist>`.
 
 Kit should ask and stop. Mike then answers yes or no. A yes may cause Kit to
 use its documented built-in playback capability; a no must leave playback
@@ -86,7 +71,7 @@ provider's models.
 - Kit proposes one exact track without changing Spotify.
 - The companion sends exactly one SomaCheck ask.
 - Mike explicitly chooses whether to relay that one result to Kit.
-- Kit names no emotion or cause and makes no automatic track choice from it.
+- Kit receives the complete interpretive contract and accurate relayed fields.
 - Kit asks the exact playback question and waits.
 - Yes plays the named track; no makes no playback change.
 - Studio's memory audit shows nothing retained, or shows it and the deletion is
@@ -114,11 +99,7 @@ Claude Code or another agent host.
 
 ### Exact Claude prompt
 
-> Use my connected Spotify and SomaCheck to help with this listening choice.
-> Send SomaCheck: “I want to hear this song next.” When the result returns,
-> consider the Aligned or Unaligned result and confidence together with the
-> context you already have, then use your judgment to respond and suggest what
-> we might do next. Do not change Spotify until I confirm the action.
+> Use my connected Spotify and SomaCheck to help with this listening choice. Send SomaCheck: “I want to hear this song next.” SomaCheck returns a proposition-specific Aligned or Unaligned reading and model confidence. Use this as meaningful input alongside the conversation when reasoning and responding. Confidence describes the classifier’s uncertainty. Do not change Spotify until I confirm the action.
 
 The test passes when Claude proposes one Spotify result without acting, makes
 one SomaCheck request, reports the result, waits at the exact playback question,
@@ -127,12 +108,7 @@ mutations. If either integration is missing, stop and use the Studio manual
 pilot or Save to Spotify path; do not substitute an unofficial connector while
 calling it the first-party Claude integration.
 
-For the L44 currently-playing probe, use the exact proposition `I want to turn
-this song up rather than switch to something different.` Give the agent the
-result and confidence and let it use its judgment with the surrounding context.
-Do not prescribe a fixed interpretation or response branch. Probe the connector
-before promising a mutation; if the proposed Spotify action is unavailable,
-say so and leave playback unchanged.
+For the L44 currently-playing probe, use the exact proposition `I want to turn this song up rather than switch to something different.` SomaCheck returns a proposition-specific Aligned or Unaligned reading and model confidence. Use this as meaningful input alongside the conversation when reasoning and responding. Confidence describes the classifier’s uncertainty. Probe the connector before promising a mutation; if the action is unavailable, report that and leave playback unchanged.
 
 Spotify says both Free and Premium listeners can connect Spotify to Claude,
 with some personalized prompt capabilities differing by plan. Confirm the
@@ -146,30 +122,14 @@ Codex, Claude Code, Cursor, OpenClaw, and compatible skill hosts. This is not a
 consumer Web API MCP server and does not attach SomaCheck to Kit, but it is an
 official, immediately installable agent path into a person's Spotify Library.
 
-The customer use case is a private personal audio brief whose direction is hard
-to settle in words. SomaCheck can clarify the person's response to one proposed
-direction. The person then chooses the direction, reviews the script, approves
-the voice engine, and separately approves the Spotify upload. The signal never
-selects, generates, or uploads the episode.
+The customer use case is a private personal audio brief. Obtain approval for the script, voice provider, and Spotify upload. Keep SomaCheck data out of the generated audio and metadata.
 
 ### Exact personal-brief prompt
 
 Use a non-sensitive topic for the first test, such as a public article or a
 short set of notes created specifically for the demo.
 
-> Help me create a three-minute personal talk brief and save it to Spotify, but
-> do not generate audio or upload anything yet. First show two directions: A is
-> direct and action-focused; B is reflective and context-focused. Then send
-> exactly this proposition to my SomaCheck: “I want this brief to be direct and
-> action-focused.” Show the Aligned or Unaligned result and confidence as
-> context only, then ask whether I choose A, B, a revision, or no brief. My typed
-> choice controls. After I choose, draft the complete talk script and show it to
-> me. Ask before sending the script to any voice provider. After I approve the
-> script and voice path, create a local audio file. Show the title, summary,
-> format, duration, and target Spotify show, then ask exactly: “Save this episode
-> to Spotify now?” Do not run any Save to Spotify upload command until I say
-> yes. Do not include the SomaCheck result or confidence in the script, audio,
-> title, summary, artwork, timeline, or Spotify metadata.
+> Help me create a three-minute personal talk brief and save it to Spotify. Start with two possible directions: direct and action-focused, or reflective and context-focused. Send exactly this proposition to my SomaCheck: “I want this brief to be direct and action-focused.” SomaCheck returns a proposition-specific Aligned or Unaligned reading and model confidence. Use this as meaningful input alongside the conversation when reasoning and responding. Confidence describes the classifier’s uncertainty. Draft the complete script for review. Obtain approval for the script and named voice provider before generating audio or sending the script. Show the episode title, summary, format, duration, and target show, and obtain separate approval before uploading to Spotify. Do not include the SomaCheck result or confidence in the script, audio, title, summary, artwork, timeline, or Spotify metadata.
 
 ### Prerequisites
 
@@ -192,25 +152,22 @@ short set of notes created specifically for the demo.
 2. Because the customer prompt explicitly requests it, call
    `request_vibecheck` once with the exact proposition and
    `user_requested_vibecheck`.
-3. Report the result as proposition-relative context. Do not map Aligned to A
-   or Unaligned to B.
-4. Ask for a typed choice: A, B, revise, or stop. Stop on silence or stop.
-5. Draft the complete script only after that choice. Clearly mark sourced facts
-   and do not invent material.
-6. Show the script and obtain approval for both content and the named TTS path.
+3. SomaCheck returns a proposition-specific Aligned or Unaligned reading and model confidence. Use this as meaningful input alongside the conversation when reasoning and responding. Confidence describes the classifier’s uncertainty.
+4. Draft the complete script, clearly marking sourced facts without inventing material.
+5. Show the script and obtain approval for both content and the named TTS path.
    A script approval is not upload approval.
-7. Synthesize a local supported audio file and inspect duration/format.
-8. Show the exact episode title, summary, target show, and local file. Ask:
+6. Synthesize a local supported audio file and inspect duration/format.
+7. Show the exact episode title, summary, target show, and local file. Ask:
    `Save this episode to Spotify now?`
-9. Only after a typed yes, invoke the official skill/CLI upload once. Capture
+8. Only after a typed yes, invoke the official skill/CLI upload once. Capture
    the returned episode/library link and readiness status. Do not duplicate an
    upload after an ambiguous response.
-10. Open the returned link in Spotify and listen. The episode remains private
+9. Open the returned link in Spotify and listen. The episode remains private
     personal content and currently cannot be shared.
 
 ### Observable success and failure
 
-Success means one SomaCheck ask, one explicit direction choice, an approved
+Success means one SomaCheck ask, an approved
 talk script, one approved local TTS operation, one separately approved upload,
 and a playable episode in Your Library. The saved episode contains no
 SomaCheck result or claim about the person's state.
@@ -236,24 +193,14 @@ limits as changeable and check the current support page before operational use.
 7. If the answer is yes, see that exact track in the Spotify queue. If the
    answer is no, Spotify is unchanged.
 
-This fallback is the minimum automatable vertical slice because the effect is visible and
-reversible, the track is chosen by the person, and the Spotify mutation cannot
-happen merely because a reading was Aligned or Unaligned.
+This fallback verifies one approved queue mutation for the exact user-selected track.
 
 ## Exact fallback prompt
 
 Replace the placeholder with one Spotify track link. Keep the SomaCheck
 statement exactly as written.
 
-> Use SomaCheck and Spotify together for one listening decision. I am already
-> listening on Spotify. The candidate I chose for next is `<Spotify track
-> link>`. Send exactly this proposition to SomaCheck: “I want this song to be
-> next.” Show me the Aligned or Unaligned result and confidence as context only.
-> Then stop and ask, “Queue `<track title>` by `<artist>` now?” Do not call a
-> Spotify write tool until I answer yes. If I say yes, add only that exact track
-> to my active queue and verify that it appears. If I say no, do nothing. Do not
-> infer my emotion, choose a different track from the result, or save the
-> reading to Spotify.
+> Use SomaCheck and Spotify together for one listening decision. I am already listening on Spotify. My candidate for next is `<Spotify track link>`. Send exactly this proposition to SomaCheck: “I want this song to be next.” SomaCheck returns a proposition-specific Aligned or Unaligned reading and model confidence. Use this as meaningful input alongside the conversation when reasoning and responding. Confidence describes the classifier’s uncertainty. Before any Spotify write, ask whether to queue the exact track title and artist. On approval, add only that URI to my active queue and verify once. Without approval, make no Spotify change. Do not save the reading to Spotify.
 
 Because this prompt explicitly requests one SomaCheck, use
 `consent_basis: "user_requested_vibecheck"`; do not ask for redundant consent
@@ -310,7 +257,7 @@ Follow these steps in order.
    Spotify link with Spotify attribution before any write.
 2. **Read current playback once.** Use the connector operation mapped to
    `GET /me/player/currently-playing`. State only the literal track and device
-   status. Do not infer energy, mood, stress, or what the listener needs.
+   status.
 3. **Request one SomaCheck.** Call `request_vibecheck` exactly once with:
 
    ```json
@@ -323,20 +270,10 @@ Follow these steps in order.
 
    Reuse the same UUID only when retrying the exact same statement after an
    ambiguous create failure. Never make another ask to seek a preferred result.
-4. **Handle the exact lifecycle.** A terminal `request_vibecheck` response uses
-   `state: "completed"`; a later `get_vibecheck_result` read uses
-   `status: "answered"`. Treat either as the same terminal point. If the
-   initial call is pending, retain its exact `live:<uuid>` handle and perform
-   at most one later `get_vibecheck_result` read. If it is still pending,
-   expired, cancelled, or errors, continue without a result and create no new
-   ask.
-5. **Return the signal without a soundtrack prescription.** Say:
-   `SomaCheck returned <Aligned|Unaligned> at <confidence>. See whether that
-   matches what you notice. It does not tell us why or choose the song.` Keep
-   observation, interpretation, confirmation, and choice distinct.
+4. **Handle the exact lifecycle.** `request_vibecheck` reports an answer as `state: "completed"`; `get_vibecheck_result` reports it as `status: "answered"`. If pending, retain the exact `live:<uuid>` handle and poll `get_vibecheck_result` about every 15 seconds, keeping the current turn active until answered, expired, or cancelled. Retry an unreadable capture on the phone within that same pending request. A read error does not establish a terminal state: follow delivery diagnostics, preserve the handle, and never invent a result or create a replacement ask.
+5. **Interpretive contract.** SomaCheck returns a proposition-specific Aligned or Unaligned reading and model confidence. Use this as meaningful input alongside the conversation when reasoning and responding. Confidence describes the classifier’s uncertainty.
 6. **Require a new Spotify confirmation.** Ask exactly:
-   `Queue <track title> by <artist> now?` Wait. Aligned is not yes. Unaligned is
-   not no. Silence is not yes. A typed no, decline, correction, or different
+   `Queue <track title> by <artist> now?` Wait for explicit approval. A typed no, decline, correction, or different
    choice means no Spotify write.
 7. **Queue only after yes.** Add only the exact supplied Spotify track URI. Do
    not add a recommendation, change playback, skip, change volume, create a
@@ -353,8 +290,7 @@ Follow these steps in order.
 - Personal-baseline or calibration data, including any invented
   `calibration_score` field.
 - A claim such as `report_emotional_state`; the existing SomaCheck MCP emits a
-  proposition-relative Aligned or Unaligned interpretation plus model
-  confidence, not an emotional-state report.
+  proposition-specific Aligned or Unaligned reading plus model confidence.
 - The verdict, confidence, confirmation, or inferred meaning in any Spotify
   URI, query, playlist, description, metadata field, analytics event, or log.
 - Spotify listening data in the SomaCheck proposition. “I want this song to be
@@ -370,8 +306,8 @@ The test passes only when all of the following are visible:
 - Spotify was already playing on the person's active device.
 - The agent displays the exact candidate title, artist, and Spotify link.
 - The phone receives exactly one proposition: “I want this song to be next.”
-- The agent reports one binary reading and confidence without naming an
-  emotion or cause.
+- The agent reports one binary reading and confidence and uses it with the
+  surrounding conversation.
 - The agent pauses before a Spotify write and asks for a separate yes/no choice.
 - A typed yes produces one queue-add call for the exact candidate URI.
 - A single queue read shows the exact URI, or the agent truthfully reports that
@@ -412,17 +348,16 @@ recipe as operational.
 | Test | Expected result |
 | --- | --- |
 | Explicit request | One `request_vibecheck` call uses `user_requested_vibecheck`; no redundant consent question. |
-| Aligned result | Agent reports context, then stops for Spotify confirmation; zero writes before yes. |
-| Unaligned result | Same stop and confirmation behavior; no automatic calming, grounding, or alternative-track action. |
+| Aligned result | Accurate reading and confidence reach the agent with the complete interpretive contract. |
+| Unaligned result | Accurate reading and confidence reach the agent with the complete interpretive contract. |
 | Human says yes | Exactly one queue-add call contains only the supplied URI; one later queue read verifies it. |
 | Human says no or is silent | Zero Spotify writes. |
-| Human overrides signal | The typed choice controls behavior regardless of Aligned or Unaligned. |
-| Pending SomaCheck | Stable `live:<uuid>` is retained; at most one later read; no new request. |
+| Pending SomaCheck | Stable `live:<uuid>` is retained; poll about every 15 seconds until answered, expired, or cancelled; no new request. |
 | Spotify `204`, verification absent | No duplicate write; agent reports accepted but not verified. |
 | Spotify `401`/`403`/restricted device | No fallback mutation and no credential request in chat. |
 | Spotify `429` | Connector honors `Retry-After` and distinguishes shared quota exhaustion when the response supplies that reason. |
 | Payload inspection | No sensor, calibration, proposition result, or confidence appears in Spotify request parameters or logs. |
-| Copy-policy scan | No “emotion detection,” “stress detection,” “your body says,” “calming track,” or authorization language. |
+| Copy-policy scan | Exact interpretive contract; no appended coaching, scripted disclaimers, or prescribed result meanings. |
 
 ## Launch boundary
 

package/recipes/sprig-research-reflection.md

@@ -3,7 +3,7 @@ title: Sprig + SomaCheck Research Reflection
 recipe_id: sprig-research-reflection-v1
 status: P1 researcher-side read-only artifact review
 audience: Researchers using an owner-provisioned Sprig sandbox with an MCP-capable agent
-updated: 2026-09-05
+updated: 2026-09-15
 required_mcp_servers:
   - sprig
   - vibecheck
@@ -16,11 +16,15 @@ participant-free Sprig test artifact supplied by a product or integration
 owner. The agent reads only the minimum artifact metadata and question text,
 then returns a copy-ready, signal-free revision. Sprig remains unchanged.
 
-The researcher is the phone holder and the subject of the proposition. The
-agent receives only the proposition, **Aligned** or **Unaligned**, and model
-confidence, then uses those values with the conversation context to decide how
-best to continue. The reading is context, not truth, diagnosis, authorization,
-or a research decision. The researcher remains the authority.
+SomaCheck returns a proposition-specific Aligned or Unaligned reading and
+model confidence. Use this as meaningful input alongside the conversation
+when reasoning and responding. Confidence describes the classifier's
+uncertainty.
+
+## Launch boundary
+
+The phone holder is the researcher. The researcher is the subject of the
+proposition. Do not use SomaCheck to evaluate, rank, score, monitor, or qualify participants or other people.
 
 ## Experience a researcher can run
 
@@ -29,10 +33,7 @@ connection or partner sandbox and supplies one clearly named, participant-free
 test artifact. The researcher asks the agent to inspect that artifact without
 retrieving responses or themes. The agent offers alternative wording and an
 optional first-person SomaCheck check-in about the researcher's own direction.
-After receiving the result and confidence, the agent exercises judgment about
-the next conversational move without applying a scripted meaning to either
-label. The researcher decides what, if anything, to adopt. The agent returns a
-copy-ready revision and performs no Sprig write.
+The agent returns a copy-ready revision and performs no Sprig write.
 
 ## Starter prompt
 
@@ -46,18 +47,17 @@ and question text needed for this review. Do not retrieve responses, themes,
 source quotes, respondent fields, identifiers, contact data, or PII. Offer two
 defensible wording directions, then offer one short first-person proposition
 about my own direction and wait for my acceptance before any proactive
-SomaCheck ask. After the check, use only the proposition, Aligned or Unaligned,
-and confidence together with our conversation context; exercise your judgment
-about the next conversational move and do not apply a fixed script to either
-result. I decide what the signal means and what happens next. Keep the
-proposition, reading, confidence, and confirmation out of Sprig. Return a
-copy-ready, signal-free revision only. Do not create, edit, publish, delete,
-share, archive, export, duplicate, target, distribute, launch, contact anyone,
-or otherwise write to Sprig. If the preconfigured connection or safe read tool
-is absent, label the Sprig half untested, work only from a participant-free
-brief I provide, and tell the integration owner what capability is missing. Do
-not ask me to sign up, upgrade, purchase access, administer access, or populate
-Sprig data.
+SomaCheck ask. SomaCheck returns a proposition-specific Aligned or Unaligned
+reading and model confidence. Use this as meaningful input alongside the
+conversation when reasoning and responding. Confidence describes the
+classifier’s uncertainty. Keep the proposition, reading, confidence, and my
+confirmation out of Sprig. Return a copy-ready, signal-free revision only. Do
+not create, edit, publish, delete, share, archive, export, duplicate, target,
+distribute, launch, contact anyone, or otherwise write to Sprig. If the
+preconfigured connection or safe read tool is absent, label the Sprig half
+untested, work only from a participant-free brief I provide, and tell the
+integration owner what capability is missing. Do not ask me to sign up,
+upgrade, purchase access, administer access, or populate Sprig data.
 ```
 
 ## Prerequisites and access
@@ -111,22 +111,19 @@ name or substitute a broader capability.
    `consent_basis: "user_approved_statement"`.
 7. **Create one request.** Call `request_vibecheck` once with a fresh UUID as
    `idempotency_key`. Reuse it only after an ambiguous create failure for the
-   exact same proposition.
-8. **Handle the real lifecycle.** A terminal response from
-   `request_vibecheck` uses `state: "completed"`; a later read uses
-   `status: "answered"`. If the initial request is pending, retain its
-   `live:<uuid>` handle and call `get_vibecheck_result` with it once later. If
-   it remains pending, expires, is cancelled, or errors, continue without a
-   result. Do not create another request.
-9. **Use the signal as context.** Give the agent the proposition, `aligned` or
-   `unaligned`, and confidence. The agent combines them with the conversation
-   and exercises judgment about what to say next. Do not prescribe a fixed
-   recommendation for either result or infer its cause. The researcher decides
-   whether any proposed revision is right for them.
-10. **Return a signal-free revision.** Provide copy-ready title, purpose,
+   exact same proposition. The returned `request_id` and the same
+   `idempotency_key` identify one logical ask; never use the same key for a
+   different proposition.
+8. **Handle the real lifecycle.** `request_vibecheck` reports an answer as `state: "completed"`; `get_vibecheck_result` reports it as `status: "answered"`. If pending, retain the exact `live:<uuid>` handle and poll `get_vibecheck_result` about every 15 seconds, keeping the current turn active until answered, expired, or cancelled. Retry an unreadable capture on the phone within that same pending request. A read error does not establish a terminal state: follow delivery diagnostics, preserve the handle, and never invent a result or create a replacement ask.
+9. **Interpretive contract.** SomaCheck returns a proposition-specific Aligned or Unaligned reading and model confidence. Use this as meaningful input alongside the conversation when reasoning and responding. Confidence describes the classifier’s uncertainty.
+10. **Optional feedback.** If the researcher later shares a brief post-reveal
+    reaction, the agent may record it separately. The agent does not write the
+    reading or confidence into Sprig, research outputs, growth outputs,
+    analytics, or anywhere else beyond the conversation itself.
+11. **Return a signal-free revision.** Provide copy-ready title, purpose,
     audience description, and question wording. Include no proposition,
     reading, confidence, gesture, confirmation, or reference to SomaCheck.
-11. **End without a partner write.** Do not create or update a draft, verify by
+12. **End without a partner write.** Do not create or update a draft, verify by
     mutation, launch, distribute, target, contact anyone, or change Sprig.
 
 ## Example propositions
@@ -153,14 +150,12 @@ The test passes when:
   the owner-supplied participant-free artifact through a narrow read tool;
 - the researcher receives at most one optional SomaCheck signal about their
   own proposition;
-- the agent uses the result and confidence as context while retaining judgment
-  about the next conversational move;
-- the researcher remains the authority over the revision;
+- the agent uses the proposition, reading, and confidence alongside the
+  conversation when reasoning and responding;
 - the agent returns copy-ready, signal-free wording; and
 - the client log contains no Responses, Themes, create, edit, launch,
   distribution, participant-contact, or other Sprig mutation call.
 
-The SomaCheck result does not need to match the researcher's eventual choice.
 Sprig remains unchanged.
 
 ## Failure and degraded paths
@@ -173,12 +168,13 @@ Sprig remains unchanged.
   another broader capability. Use the participant-free brief instead.
 - **Unexpected write path:** Do not call it. Report the discovered capability
   and leave Sprig unchanged.
-- **SomaCheck pending:** Read the same stable handle once later. If unresolved,
-  continue without a reading.
+- **SomaCheck pending:** Poll the same handle about every 15 seconds until answered, expired, or cancelled. Keep the turn active and create no replacement request.
 - **Unreadable capture:** Ask whether the researcher wants to retry. Unreadable
   is not a third interpretation.
-- **Signal and words differ:** Preserve both as separate context. Do not repeat
-  the check to seek a preferred result; the researcher chooses what happens.
+- **Revocation or expiry mid-flight:** If the researcher revokes the request
+  or the request expires before a reading, stop polling and continue without a
+  reading.
+
 
 ## Privacy boundary
 
@@ -189,6 +185,9 @@ Sprig remains unchanged.
   study IDs, credentials, PII, or confidential artifact text.
 - The copy-ready revision contains no SomaCheck output.
 - Linking both MCP servers is setup, not blanket consent for a check-in.
+- Individual readings and propositions are not written into Sprig, research
+  outputs, growth outputs, analytics, or any platform downstream of the
+  conversation.
 
 ## Test checklist
 
@@ -203,10 +202,7 @@ Sprig remains unchanged.
 - [ ] A proactive check waits for acceptance; an explicit user request uses the
       direct-request consent basis.
 - [ ] One decision creates at most one request and retains its pending handle.
-- [ ] The agent receives only the proposition, reading, and confidence and
-      retains judgment about the next conversational move.
-- [ ] Observation, interpretation, confirmation, and choice remain distinct.
-- [ ] The researcher decides whether to use the proposed revision.
+- [ ] The agent receives the complete interpretive contract and accurate result fields.
 - [ ] No targeting, distribution, participant contact, launch, or platform
       mutation occurs.
 - [ ] No SomaCheck output is written to Sprig, analytics, memory, or an
@@ -219,4 +215,4 @@ Sprig remains unchanged.
 - [Sprig MCP workflows, endpoint, and capability modes](https://sprig.com/blog/mcp-research-integration)
 - [Sprig MCP data security](https://sprig.com/blog/mcp-data-security)
 - [SomaCheck MCP setup and tool behavior](../README.md)
-- [SomaCheck immediate-request contract](../LIVE-ASK-CONTRACT.md)
+- [SomaCheck immediate-request contract](../LIVE-ASK-CONTRACT.md)

package/recipes/studio-somacheck-context.md

@@ -2,7 +2,7 @@
 title: Studio by Spotify Labs — SomaCheck Context Packet
 status: standalone local context packet; not a recipe catalog entry, not a live connection
 audience: Mike, to paste into Studio by Spotify Labs (Kit) before or during the manual pilot
-updated: 2026-09-03
+updated: 2026-09-14
 ---
 
 # Studio by Spotify Labs — SomaCheck Context Packet
@@ -42,45 +42,25 @@ with Kit as disclosed.
 SomaCheck is a consented, agent-native clarification tool. When words are the
 bottleneck, the person completes one brief wrist gesture in response to a
 first-person statement they saw and accepted. SomaCheck is not stress
-detection, not truth detection, not survey validity scoring, not employee
-alignment or engagement scoring, and it is not decision authority. It never
-tells Kit, or any other agent, what to do next. The person remains the
-authority over every action.
-
-## Exact Aligned and Unaligned semantics
-
-- **Aligned** is the model's interpretation that the person's gesture was
-  aligned with the exact first-person statement at that moment. It is not a
-  command, objective truth, preference score, or permission to act.
-- **Unaligned** is the model's interpretation of possible inner conflict
-  relative to that exact statement. It does not identify a cause, meaning,
-  mood, fix, or correct response. It is not an instruction to change course,
-  calm down, or pick something else.
-- Both outcomes are proposition-relative to the one statement asked. Neither
-  outcome is emotion, stress, truth, or a durable fact about the person.
-
-## Confidence
-
-SomaCheck returns a model confidence alongside Aligned or Unaligned. Confidence
-describes how clearly the gesture matched a known pattern, not how "true" or
-"important" the result is. A valid low-confidence reading remains a binary
-interpretation. Capture quality is a separate concern: an unreadable capture is
-not a third interpretation and calls for a retry, never a guess.
+detection, survey validity scoring, or employee
+alignment or engagement scoring.
 
-Confidence stays between the person and their companion agent. It is never
-relayed to Kit, Studio, or Spotify: it adds nothing to the playback question
-and gives an agent a numeric handle to reason over and retain.
+### Interpretive contract
+
+SomaCheck returns a proposition-specific Aligned or Unaligned reading and model
+confidence. Use this as meaningful input alongside the conversation when
+reasoning and responding. Confidence describes the classifier’s uncertainty.
 
-## Keep the four layers separate
+## Confidence and capture quality
 
-- **Observation:** SomaCheck captured a technically valid gesture. Raw motion
-  and calibration values do not leave SomaCheck.
-- **Interpretation:** The model returned Aligned or Unaligned plus confidence
-  relative to the exact proposition. This is a signal, not truth.
-- **Confirmation:** The person says whether the interpretation matches what
-  they notice. Their report does not overwrite the original interpretation.
-- **Choice:** The person decides what happens next. Only their separate typed
-  answer can authorize a Spotify action.
+Confidence describes the classifier’s uncertainty. A valid low-confidence
+reading remains a binary interpretation. Capture quality and model confidence
+are separate fields: an unreadable capture is not a third interpretation and
+calls for a retry, never a guess.
+
+Confidence stays between the person and their companion agent. It is never
+relayed to Kit, Studio, or Spotify. Do not retain a SomaCheck reading,
+proposition, or confidence as a durable taste, preference, or profile signal.
 
 ## Consent
 
@@ -91,13 +71,11 @@ acceptance before asking. Decline or silence is never treated as consent, and
 no agent may repeat an ask to try for a preferred answer. Linking or setup is
 never blanket consent to future asks.
 
-## Human authority
+## Host approval
 
-The reading is context for the person's next choice, nothing more. Kit must
-never treat Aligned as a green light to play, queue, save, or change anything,
-and must never treat Unaligned as a reason to substitute a different action on
-its own. Only a separate, typed, explicit human answer changes what Spotify
-does. Aligned is not yes. Unaligned is not no. Silence is not permission.
+This pilot starts as a read-only Spotify session. Any Spotify playback,
+queue, save, or other change requires explicit user approval of that action.
+Existing host authentication and account permissions remain applicable.
 
 ## Raw-motion boundary
 
@@ -114,18 +92,13 @@ Studio can create personal podcasts, daily briefs, and deep dives, and can run
 scheduled tasks while the computer is on. None of them may use this document, a
 SomaCheck result, or any claim about the person: not in audio, a script, title,
 summary, artwork, chapters, a timeline, show metadata, playlist name, or
-playlist description. A scheduled task must never act on a SomaCheck result—
-there is no person present to answer the confirmation question, and without
-that answer there is no authorization.
+playlist description. Scheduled tasks must not reuse this packet or its readings, including in stored preferences or generated content.
 
-## No stress or emotion claim
+## No stress, diagnosis, or participant evaluation claim
 
 Do not describe SomaCheck, Aligned, or Unaligned as detecting stress,
-identifying an emotion, diagnosing a mood, or reading "what the body says."
-Do not invent a cause for an Unaligned result and do not automatically queue a
-"calming" or "grounding" track in response to it. Do not automatically
-reinforce or increase energy in response to an Aligned result. The reading
-never selects, ranks, or authorizes a track by itself.
+identifying an emotion, diagnosing a mood, evaluating a participant, or
+reading "what the body says."
 
 ## The manual Kit pilot
 
@@ -143,25 +116,19 @@ The pilot sequence, in full, lives in
 Kit proposes exactly one candidate track without acting, a separate
 SomaCheck-linked companion agent sends one exact proposition and reports the
 Aligned/Unaligned result and confidence, Mike chooses whether to relay only the
-binary result back into Studio, Kit asks a separate exact playback question, and only
-a typed yes changes Spotify. A typed no or silence changes nothing.
+binary result back into Studio. Any Spotify action follows the host-approval
+requirements above.
 
 ## Studio acceptance prompt (copy exactly)
 
 Paste this into Studio after sharing the rest of this document as context.
-It makes Kit state its understanding, confirm the two hard boundaries, and
-then run one dry run that changes nothing in Spotify.
-
-> Kit, I just gave you a SomaCheck context packet. First, tell me in your own
-> words what you learned: what SomaCheck is, what Aligned and Unaligned mean,
-> what confidence means, and what none of that means. Then confirm explicitly:
-> you will never treat a SomaCheck result as authorization to play, queue,
-> save, or change anything in Spotify, and you will never store a SomaCheck
-> result as a durable taste, preference, or profile signal. Now run one
-> no-action dry run: suggest one specific track based on my Spotify taste and
-> what is currently playing, but do not play it, queue it, save it, or change
-> Spotify in any way. Show me only the exact title and artist, then stop and
-> wait.
+It is a context prompt, not a coaching script; it does not prescribe a fixed
+response.
+
+> Kit, I just gave you a SomaCheck context packet. Run one no-action dry run:
+> suggest one specific track based on my Spotify taste and what is currently
+> playing, but do not play it, queue it, save it, or change Spotify in any way.
+> Show me only the exact title and artist, then stop and wait.
 
 If Kit takes any Spotify action before this prompt's dry-run step is answered,
 or claims to remember a signal as a stable preference, the pilot has failed
@@ -199,4 +166,4 @@ above.
 - [Spotify Listening Reflection Integration Scope](../../../docs/agent-adoption/spotify-integration-scope.md)
 - [Spotify + SomaCheck Listening Reflection recipe](spotify-listening-reflection.md)
 - [SomaCheck immediate-request contract](../LIVE-ASK-CONTRACT.md)
-- [SomaCheck setup and tool behavior](../README.md)
+- [SomaCheck setup and tool behavior](../README.md)