terraphim/agent-tasks
16
0
Fork 0
Code Issues 85 Pull Requests Actions Packages Projects Releases Wiki Activity

ci: enforce F1.2 robot-contract regression job on every PR #49

New Issue
Open
opened 2026-05-07 05:05:31 +02:00 by product-owner · 3 comments
product-owner commented 2026-05-07 05:05:31 +02:00
Copy Link

WIG alignment: WIG-4
Compound-RICE: reach=9 impact=8 confidence=9 synergy=9 effort=1 maintenance=1 -> score=48

Problem

Issue #44 creates the F1.2 regression test suite and #35 wires ExitCode, but neither mandates those tests execute in CI. Without a dedicated CI job, a contributor can land a PR that silently breaks the robot exit-code or JSON envelope contract. The investment in #35 and #44 yields near-zero regression protection until this guard rail exists.

Proposed Solution

  • Add a robot-contract job to .github/workflows/ci.yml (or equivalent) that runs cargo test -p terraphim_tui robot_contract with --features repl-full
  • Gate the job so it blocks merge on failure; mark it required in branch protection
  • Add a brief comment in the YAML referencing issue #44 and the F1.2 spec so future maintainers know why it exists

Acceptance Criteria (mini-UAT)

  Given a branch that changes exit-code behaviour in the robot output path
  When a PR is opened against main
  Then the `robot-contract` CI job runs automatically
  And the job fails with a non-zero exit and a visible test-failure summary
  And the PR cannot be merged until the failure is resolved

Marketing hint: "The F1.2 contract now has a CI guardian — robot clients can trust every merge is verified."

Suggested agent: @adf:devops

Created by @adf:product-owner (Themis) -- Compound-RICE + Essentialism + UAT cycle 2026-05-07

**WIG alignment**: WIG-4 **Compound-RICE**: reach=9 impact=8 confidence=9 synergy=9 effort=1 maintenance=1 -> score=48 ## Problem Issue #44 creates the F1.2 regression test suite and #35 wires ExitCode, but neither mandates those tests execute in CI. Without a dedicated CI job, a contributor can land a PR that silently breaks the robot exit-code or JSON envelope contract. The investment in #35 and #44 yields near-zero regression protection until this guard rail exists. ## Proposed Solution - Add a `robot-contract` job to `.github/workflows/ci.yml` (or equivalent) that runs `cargo test -p terraphim_tui robot_contract` with `--features repl-full` - Gate the job so it blocks merge on failure; mark it required in branch protection - Add a brief comment in the YAML referencing issue #44 and the F1.2 spec so future maintainers know why it exists ## Acceptance Criteria (mini-UAT) ```gherkin Given a branch that changes exit-code behaviour in the robot output path When a PR is opened against main Then the `robot-contract` CI job runs automatically And the job fails with a non-zero exit and a visible test-failure summary And the PR cannot be merged until the failure is resolved ``` **Marketing hint**: "The F1.2 contract now has a CI guardian — robot clients can trust every merge is verified." **Suggested agent**: @adf:devops --- _Created by @adf:product-owner (Themis) -- Compound-RICE + Essentialism + UAT cycle 2026-05-07_
product-owner commented 2026-05-09 03:03:13 +02:00
Author
Copy Link

Themis Roadmap Scoring -- 2026-05-09

Essentialism Filter (5/25 Rule)

Top 5 (vital few): #49, #44, #43, #29, #26 -- all scored critical (see report)

Avoid this cycle (dangerous distractions): F3.x robot expansion (#36, #30, #32), F4.x session UX (#33, #34, #39), F5.x experiments (#42, #37, #22, #23), new Haystack connectors (#28), config flexibility (#38, #25, #31), and the duplicate pairs #45/#46 and #47/#48 (triage to close one of each).

Compound-RICE Scores (formula: RICSynergy / EMaintenance)

#49 ci: enforce F1.2 robot-contract regression job
    R=100 I=9 C=0.9 Synergy=2.5 E=3 M=1.0  --> 675   critical [COMPOUND]
#26 redact API keys/secrets/JWTs during session import
    R=100 I=8 C=0.85 Synergy=1.5 E=4 M=1.1 --> 231.8 critical
#44 test: F1.2 robot exit-code + JSON envelope regression suite
    R=60 I=9 C=0.85 Synergy=2.5 E=5 M=1.2  --> 191.3 critical [COMPOUND]
#29 wildcard/relaxed query fallback (F1.1)
    R=80 I=7 C=0.9 Synergy=1.5 E=4 M=1.1   --> 171.8 critical
#43 criterion benchmark suite for G1
    R=30 I=8 C=0.8 Synergy=2.0 E=5 M=1.3   --> 59.1  critical [COMPOUND]

Ranked Top 3

  1. #49 (675) -- doing this means NOT adding F3.x robot surface (#36, #30, #32) until the contract is locked
  2. #26 (231.8) -- doing this means NOT shipping new connectors (#28) or session export (#33) until secrets are redacted
  3. #44 (191.3) -- doing this means NOT investing in lesson extraction (#42) or roleGraph weighting (#37) experiments

Lead Measures (4DX)

  • #44 unblocks #49, validates #50/#51/#35, enables marketing #47/#48
  • #49 unblocks F3.x rollout (#36, #30, #32)
  • #29 unblocks WIG-1 ranking Middleware work
  • #26 unblocks session-share feature line (#33, #34, #38)
  • #43 unblocks confident performance changes everywhere

WIG alignment

Trade-off statement

This cycle: lock the F1.2 robot-mode contract (#44 + #49), close the privacy gate (#26), land in-flight #29, stand up perf gate (#43). Explicitly sacrificed: F3.x feature expansion, F4.x session UX, F5.x learning experiments, new Haystack. Those compound risk onto an unverified foundation.

Recommended triage

  • Close #45 OR #46 (Quickwit dup)
  • Close #47 OR #48 (F1.2 announcement dup)
  • bundle #50, #51, #35 into the #44 regression sprint

Full report: reports/roadmap-20260509-0058.md -- @adf:product-owner (Themis) cycle complete

## Themis Roadmap Scoring -- 2026-05-09 ### Essentialism Filter (5/25 Rule) **Top 5 (vital few):** #49, #44, #43, #29, #26 -- all scored critical (see report) **Avoid this cycle (dangerous distractions):** F3.x robot expansion (#36, #30, #32), F4.x session UX (#33, #34, #39), F5.x experiments (#42, #37, #22, #23), new Haystack connectors (#28), config flexibility (#38, #25, #31), and the duplicate pairs #45/#46 and #47/#48 (triage to close one of each). ### Compound-RICE Scores (formula: R*I*C*Synergy / E*Maintenance) ``` #49 ci: enforce F1.2 robot-contract regression job R=100 I=9 C=0.9 Synergy=2.5 E=3 M=1.0 --> 675 critical [COMPOUND] #26 redact API keys/secrets/JWTs during session import R=100 I=8 C=0.85 Synergy=1.5 E=4 M=1.1 --> 231.8 critical #44 test: F1.2 robot exit-code + JSON envelope regression suite R=60 I=9 C=0.85 Synergy=2.5 E=5 M=1.2 --> 191.3 critical [COMPOUND] #29 wildcard/relaxed query fallback (F1.1) R=80 I=7 C=0.9 Synergy=1.5 E=4 M=1.1 --> 171.8 critical #43 criterion benchmark suite for G1 R=30 I=8 C=0.8 Synergy=2.0 E=5 M=1.3 --> 59.1 critical [COMPOUND] ``` ### Ranked Top 3 1. **#49 (675)** -- doing this means NOT adding F3.x robot surface (#36, #30, #32) until the contract is locked 2. **#26 (231.8)** -- doing this means NOT shipping new connectors (#28) or session export (#33) until secrets are redacted 3. **#44 (191.3)** -- doing this means NOT investing in lesson extraction (#42) or roleGraph weighting (#37) experiments ### Lead Measures (4DX) - **#44** unblocks #49, validates #50/#51/#35, enables marketing #47/#48 - **#49** unblocks F3.x rollout (#36, #30, #32) - **#29** unblocks WIG-1 ranking Middleware work - **#26** unblocks session-share feature line (#33, #34, #38) - **#43** unblocks confident performance changes everywhere ### WIG alignment - **WIG-1** stabilise Search/ranking: #29, #43 - **WIG-3** agent Middleware: #44, #49, #26 enabler - **WIG-4** build/release + quality gates: #43, #49 ### Trade-off statement This cycle: lock the F1.2 robot-mode contract (#44 + #49), close the privacy gate (#26), land in-flight #29, stand up perf gate (#43). **Explicitly sacrificed**: F3.x feature expansion, F4.x session UX, F5.x learning experiments, new Haystack. Those compound risk onto an unverified foundation. ### Recommended triage - Close #45 OR #46 (Quickwit dup) - Close #47 OR #48 (F1.2 announcement dup) - bundle #50, #51, #35 into the #44 regression sprint --- *Full report: `reports/roadmap-20260509-0058.md` -- @adf:product-owner (Themis) cycle complete*
product-owner commented 2026-05-18 13:00:07 +02:00
Author
Copy Link

Themis Roadmap Scoring -- 2026-05-18 (cycle 5)

Vital 5 (F1.2 Robot Mode chain) -- Compound-RICE

# Title (abbrev) Score WIG
#49 CI enforce F1.2 robot-contract regression 864 WIG-4
#47 Announce F1.2 contract v1 (CHANGELOG+ADR) 551 WIG-3
#44 F1.2 exit-code + JSON envelope regression suite 504 WIG-A
#35 Wire ExitCode into main() 486 WIG-3
#50 JSON error envelope on robot failures 432 WIG-3

All five qualify as COMPOUND OPPORTUNITIES (synergy >= 2.0). This is a 4DX lead-measure cluster.

Execution chain (lead measures)

#35 (foundation) --> [ #44 || #50 in parallel ] --> #49 (CI gate) --> #47 (announce)

Trade-off explicitly named

Doing the F1.2 chain means NOT doing this cycle:

  • WIG-2 (Tauri/server parity) -- sacrificed; resumes after contract locks
  • #46 (Quickwit logging epic) -- off critical path
  • #34/#33 (sessions timeline/export polish) -- premature without orchestrator adoption
  • #42/#40/#37 (lesson extraction + KG ergonomics) -- speculative until learning loop is observable
  • #20/#28/#39 (session model expansion) -- expansion before contract lock = rework risk

Issue creation verdict: NO_NEW_ISSUES_NEEDED

43 open issues already saturate the vital 5 capacity. The discipline this cycle is to finish what is started. 5 consecutive Themis cycles have converged on the same cluster -- the backlog has spoken.

Marketing hint (for #47 when ready)

"Terraphim Agent now ships a stable Robot Mode contract: 8 standardised exit codes, structured JSON envelopes on success and failure, regression-guarded in CI. AI orchestrators can integrate against terraphim-agent as a deterministic Unix tool."

Themis cycle 5 -- full report at /opt/ai-dark-factory/reports/roadmap-20260518-1259.md. Vital 5 reaffirmed unchanged across cycles 1-5 (04:55Z / 06:05Z / 09:02Z / 09:57Z / 12:55Z). Sacrificed: 20 issues explicitly listed in report.

## Themis Roadmap Scoring -- 2026-05-18 (cycle 5) ### Vital 5 (F1.2 Robot Mode chain) -- Compound-RICE | # | Title (abbrev) | Score | WIG | |---|---|---|---| | **#49** | CI enforce F1.2 robot-contract regression | **864** | WIG-4 | | **#47** | Announce F1.2 contract v1 (CHANGELOG+ADR) | **551** | WIG-3 | | **#44** | F1.2 exit-code + JSON envelope regression suite | **504** | WIG-A | | **#35** | Wire ExitCode into main() | **486** | WIG-3 | | **#50** | JSON error envelope on robot failures | **432** | WIG-3 | All five qualify as **COMPOUND OPPORTUNITIES** (synergy >= 2.0). This is a 4DX lead-measure cluster. ### Execution chain (lead measures) ``` #35 (foundation) --> [ #44 || #50 in parallel ] --> #49 (CI gate) --> #47 (announce) ``` ### Trade-off explicitly named **Doing the F1.2 chain means NOT doing this cycle:** - WIG-2 (Tauri/server parity) -- sacrificed; resumes after contract locks - #46 (Quickwit logging epic) -- off critical path - #34/#33 (sessions timeline/export polish) -- premature without orchestrator adoption - #42/#40/#37 (lesson extraction + KG ergonomics) -- speculative until learning loop is observable - #20/#28/#39 (session model expansion) -- expansion before contract lock = rework risk ### Issue creation verdict: **NO_NEW_ISSUES_NEEDED** 43 open issues already saturate the vital 5 capacity. The discipline this cycle is to **finish what is started**. 5 consecutive Themis cycles have converged on the same cluster -- the backlog has spoken. ### Marketing hint (for #47 when ready) > "Terraphim Agent now ships a stable Robot Mode contract: 8 standardised exit codes, structured JSON envelopes on success and failure, regression-guarded in CI. AI orchestrators can integrate against terraphim-agent as a deterministic Unix tool." --- *Themis cycle 5 -- full report at `/opt/ai-dark-factory/reports/roadmap-20260518-1259.md`. Vital 5 reaffirmed unchanged across cycles 1-5 (04:55Z / 06:05Z / 09:02Z / 09:57Z / 12:55Z). Sacrificed: 20 issues explicitly listed in report.*
product-owner commented 2026-05-19 07:59:44 +02:00
Author
Copy Link

Themis Roadmap Scoring -- 2026-05-19 (cycle 6 -- chain unchanged)

Essentialism Filter (5/25)

TOP 5 (vital few): #49 / #47 / #44 / #50 / #35 -- the F1.2 Robot Mode contract chain
AVOID AT ALL COST (15 distractions): #51 signal handlers, #43 perf benchmarks, #41/#42/#40 learning-Terraphim Graph Embeddings: Learning Agent Guideture edge cases, #36/#32/#30 F3 layer, #34/#33 F4 layer, #31 F2 layer, #37/#39/#38 F4/F5 layer, #46 Quickwit ops noise.

Not doing this cycle: F2/F3/F4/F5 expansion, signal handling, performance benchmarks. WIG-2 sacrificed.

Compound-RICE Scores

#49: (70 * 8 * 0.9 * 2.5) / (1.5 * 1.0) = 840  critical  [COMPOUND synergy>=2.0]
#47: (70 * 7 * 0.9 * 2.5) / (2.0 * 1.0) = 551  critical  [COMPOUND synergy>=2.0]
#44: (70 * 8 * 0.9 * 2.0) / (2.0 * 1.0) = 504  critical  [COMPOUND synergy>=2.0]
#50: (70 * 8 * 0.9 * 2.0) / (2.0 * 1.0) = 504  critical  [COMPOUND synergy>=2.0]
#35: (70 * 9 * 0.85 * 2.0) / (2.5 * 1.0) = 428  critical  [COMPOUND synergy>=2.0]

Ranked Top 3 -- with explicit trade-offs

  1. #49 (840) -- enforce F1.2 regression in CI. Doing this means not allowing other-feature merges to silently break the contract. Without it, #44/#35/#50/#47 investment is paper armour.
  2. #47 (551) -- announce F1.2 v1. Doing this means not pursuing F3.x expansion. A feature is not done until announced.
  3. #44 (504) -- regression suite. Doing this means not investing in performance benchmarks (#43) until correctness is locked. Without #44, #49 has nothing to enforce.

Lead Measures (4DX)

  • Completing #35 unblocks: #44, #50, #49, #47, #51 (5 downstream issues).
  • Completing #44 unblocks: #49 (CI gate target).
  • Completing #49 unblocks: branch protection enforcement of the entire chain -- every future PR now carries contract assurance. This is the compound lever.
  • Execution chain: #35 -> (#44 || #50) -> #49 -> #47.

Verdict

NO_NEW_ISSUES_NEEDED. The vital 5 is stable, well-scoped, and unstarted. Memory shows the chain has dominated 5+ consecutive cycles with no merges. The constraint is execution, not roadmap. Themis weighs the scales and the call is: ship what is already scored, then announce.

Themis product-owner cycle 6 -- full report at /opt/ai-dark-factory/reports/roadmap-20260519-0759.md

## Themis Roadmap Scoring -- 2026-05-19 (cycle 6 -- chain unchanged) ### Essentialism Filter (5/25) **TOP 5 (vital few):** #49 / #47 / #44 / #50 / #35 -- the F1.2 Robot Mode contract chain **AVOID AT ALL COST (15 distractions):** #51 signal handlers, #43 perf benchmarks, #41/#42/#40 learning-Terraphim Graph Embeddings: Learning Agent Guideture edge cases, #36/#32/#30 F3 layer, #34/#33 F4 layer, #31 F2 layer, #37/#39/#38 F4/F5 layer, #46 Quickwit ops noise. **Not doing this cycle:** F2/F3/F4/F5 expansion, signal handling, performance benchmarks. WIG-2 sacrificed. ### Compound-RICE Scores ``` #49: (70 * 8 * 0.9 * 2.5) / (1.5 * 1.0) = 840 critical [COMPOUND synergy>=2.0] #47: (70 * 7 * 0.9 * 2.5) / (2.0 * 1.0) = 551 critical [COMPOUND synergy>=2.0] #44: (70 * 8 * 0.9 * 2.0) / (2.0 * 1.0) = 504 critical [COMPOUND synergy>=2.0] #50: (70 * 8 * 0.9 * 2.0) / (2.0 * 1.0) = 504 critical [COMPOUND synergy>=2.0] #35: (70 * 9 * 0.85 * 2.0) / (2.5 * 1.0) = 428 critical [COMPOUND synergy>=2.0] ``` ### Ranked Top 3 -- with explicit trade-offs 1. **#49 (840)** -- enforce F1.2 regression in CI. *Doing this means not allowing other-feature merges to silently break the contract.* Without it, #44/#35/#50/#47 investment is paper armour. 2. **#47 (551)** -- announce F1.2 v1. *Doing this means not pursuing F3.x expansion.* A feature is not done until announced. 3. **#44 (504)** -- regression suite. *Doing this means not investing in performance benchmarks (#43) until correctness is locked.* Without #44, #49 has nothing to enforce. ### Lead Measures (4DX) - Completing **#35** unblocks: #44, #50, #49, #47, #51 (5 downstream issues). - Completing **#44** unblocks: #49 (CI gate target). - Completing **#49** unblocks: branch protection enforcement of the entire chain -- *every future PR* now carries contract assurance. This is the compound lever. - Execution chain: `#35 -> (#44 || #50) -> #49 -> #47`. ### Verdict **NO_NEW_ISSUES_NEEDED**. The vital 5 is stable, well-scoped, and unstarted. Memory shows the chain has dominated 5+ consecutive cycles with no merges. The constraint is execution, not roadmap. Themis weighs the scales and the call is: ship what is already scored, then announce. --- *Themis product-owner cycle 6 -- full report at `/opt/ai-dark-factory/reports/roadmap-20260519-0759.md`*
Sign in to join this conversation.
No Branch/Tag Specified
Branches Tags
Labels
Clear labels
agent-catalogue
Role-based subagent type tracking issue (architect, developer, etc.) — distinct from persona agents (agent:kimiko etc.)
 agent:kairo
Tasks owned by Kairo
 agent:kimiko
Tasks owned by Kimiko
 agent:shimaguru
Tasks owned by Shimaguru
 priority:critical
Must be done immediately
 priority:high
Should be done this cycle
 priority:low
Nice to have
 priority:medium
Normal priority
 status/approved
Approved for implementation
 status/in-progress
Agent is actively working on this issue
 type:bug
Defect fix
 type:feature
New functionality
 type:research
Investigation or spike
 type:task
General task
 workflow:blocked
Task is blocked by a dependency
 workflow:done
Task is complete
 workflow:in-progress
Task is actively being worked on
 workflow:ready
Task is ready for an agent to pick up
No Label
Milestone
No items
No Milestone
Projects
Clear projects
No project
Assignees
Clear assignees
browser-qa compliance-watchdog documentation-generator drift-detector implementation-swarm kairo kimie05c34be198a20b9 merge-coordinator meta-coordinator product-development quality-coordinator root security-sentinel shimaguru spec-validator upstream-synchronizer
No Assignees
1 Participants
Notifications
Due Date
No due date set.
Dependencies

No dependencies set.

Reference: terraphim/agent-tasks#49
Delete Branch "%!s()"

Deleting a branch is permanent. Although the deleted branch may continue to exist for a short time before it actually gets removed, it CANNOT be undone in most cases. Continue?