feat(agents): add virtual provider backends config and round-robin routing #157

Merged

CoreRasurae merged 2 commits from feature/m1-virtual-provider-round-robin into develop 2026-10-19 21:15:07 +00:00

Conversation 1 Commits 2 Files changed 20 +2726 -9

CoreRasurae commented 2026-10-16 16:51:11 +00:00

Member

Copy link

Summary

Implements issue #155 (the first increment of Epic #154, ADR-2038): an optional backends list (>= 2 entries) and strategy field (currently only round_robin) on type: llm agents, so a single logical agent can pool several backend credential sets — e.g. multiple OAuth subscriptions for the same provider — and round-robin requests across them to multiply effective capacity.

Key design points (full rationale in ADR-2038, status: approved):

• VirtualChatModel is a genuine BaseChatModel subclass (same precedent as ReasoningChatModel, ADR-2036) — Liskov-substitutable, no changes needed anywhere else in the agent/tool-loop/graph dispatch stack.
• Each backend is built via the existing, unmodified per-provider construction/credential-validation logic; credentials are never merged into the shared provider-keyed credentials dict.
• Backend selection happens exactly once per external invocation; retries (ADR-2032) and the first-chunk-only streaming retry are pinned to that same selected backend — LLMAgent._retry_ainvoke/_run_pruning_pass/stream_message required no changes.
• reasoning_feedback (ADR-2036) is read once at the virtual-provider level and applied identically to every backend, so reasoning-aware backend switching still round-trips reasoning_content correctly.
• Also fixes a pre-existing Behave test-isolation bug found while adding these scenarios: reasoning_provider_routing_steps.py's _reasoning_unavailable context flag wasn't reset between scenarios, silently leaking into later scenarios that reuse its shared "the chat model is constructed" step.

Test plan

• nox -s lint — green
• nox -s typecheck — green (Pyright strict, zero suppressions)
• nox -s unit_tests — green, full suite (155 features / 3144 scenarios)
• nox -s coverage_report — 96.9% (threshold 96.5%)
• nox -s security_scan — green (bandit + semgrep)
• nox -s dead_code — green (vulture)
• nox -s integration_tests -- robot/virtual_provider_routing.robot — green (Executor → LLMAgent → VirtualChatModel end-to-end through a real tool-call loop, >= 2 backends)
• ASV benchmark added (benchmarks/virtual_provider_benchmark.py) for the per-invocation backend-selection hot path
• nox -s docs — mkdocs build green with the new guide page
• docs/index.md §4.4/§4.4.7 updated, Version bumped 1.7.0 → 1.8.0, §21.1 revision-history row added
• examples/packages/virtual-provider-pooled-agent.yaml added
• docs/guides/virtual-provider-backend-routing.md added (practical guide, mirroring the reasoning-aware-llm-agents guide)
• CHANGELOG.md updated

Closes #155

## Summary Implements issue #155 (the first increment of Epic #154, ADR-2038): an optional `backends` list (>= 2 entries) and `strategy` field (currently only `round_robin`) on `type: llm` agents, so a single logical agent can pool several backend credential sets — e.g. multiple OAuth subscriptions for the same provider — and round-robin requests across them to multiply effective capacity. Key design points (full rationale in ADR-2038, status: approved): - `VirtualChatModel` is a genuine `BaseChatModel` subclass (same precedent as `ReasoningChatModel`, ADR-2036) — Liskov-substitutable, no changes needed anywhere else in the agent/tool-loop/graph dispatch stack. - Each backend is built via the existing, unmodified per-provider construction/credential-validation logic; credentials are never merged into the shared provider-keyed credentials dict. - Backend selection happens exactly once per external invocation; retries (ADR-2032) and the first-chunk-only streaming retry are pinned to that same selected backend — `LLMAgent._retry_ainvoke`/`_run_pruning_pass`/`stream_message` required **no changes**. - `reasoning_feedback` (ADR-2036) is read once at the virtual-provider level and applied identically to every backend, so reasoning-aware backend switching still round-trips `reasoning_content` correctly. - Also fixes a pre-existing Behave test-isolation bug found while adding these scenarios: `reasoning_provider_routing_steps.py`'s `_reasoning_unavailable` context flag wasn't reset between scenarios, silently leaking into later scenarios that reuse its shared "the chat model is constructed" step. ## Test plan - [x] `nox -s lint` — green - [x] `nox -s typecheck` — green (Pyright strict, zero suppressions) - [x] `nox -s unit_tests` — green, full suite (155 features / 3144 scenarios) - [x] `nox -s coverage_report` — 96.9% (threshold 96.5%) - [x] `nox -s security_scan` — green (bandit + semgrep) - [x] `nox -s dead_code` — green (vulture) - [x] `nox -s integration_tests -- robot/virtual_provider_routing.robot` — green (Executor → LLMAgent → VirtualChatModel end-to-end through a real tool-call loop, >= 2 backends) - [x] ASV benchmark added (`benchmarks/virtual_provider_benchmark.py`) for the per-invocation backend-selection hot path - [x] `nox -s docs` — mkdocs build green with the new guide page - [x] `docs/index.md` §4.4/§4.4.7 updated, Version bumped 1.7.0 → 1.8.0, §21.1 revision-history row added - [x] `examples/packages/virtual-provider-pooled-agent.yaml` added - [x] `docs/guides/virtual-provider-backend-routing.md` added (practical guide, mirroring the reasoning-aware-llm-agents guide) - [x] CHANGELOG.md updated Closes #155

CoreRasurae added this to the v2.1.0 milestone 2026-10-16 16:51:11 +00:00

CoreRasurae added the

Type

Feature

label 2026-10-16 16:51:11 +00:00

CoreRasurae added 1 commit 2026-10-16 16:51:11 +00:00

feat(agents): add virtual provider backends config and round-robin routing ... 9a16cc5955

Adds an optional `backends` list (>= 2 entries) and `strategy` field
(default and currently only supported value: `round_robin`) to `type: llm`
agent configuration, so a single logical agent can pool several backend
credential sets -- e.g. multiple OAuth subscriptions for the same provider
-- and forward each invocation to exactly one of them. This solves the
"single provider/model/credential-set per agent" limitation of the shared,
provider-keyed credentials dict, which cannot hold two distinct credential
sets for the same provider name.

Design is recorded in ADR-2038. Key decisions carried into this
implementation:
- `VirtualChatModel` is a genuine `BaseChatModel` subclass (same precedent
  as `ReasoningChatModel`, ADR-2036), overriding only
  `_generate`/`_agenerate`/`_stream`/`_astream`/`_llm_type` so it is
  Liskov-substitutable for any other chat model this project constructs.
- Each backend is built via the existing, unmodified per-provider
  construction/credential-validation logic
  (`llm_client._build_from_credentials`) -- no parallel validation path.
- Backend selection happens exactly once per external invocation; retries
  (ADR-2032) and the first-chunk-only streaming retry are performed
  internally against that same selected backend, never re-selecting
  mid-retry or mid-stream -- otherwise a retry would silently bounce
  between backends, which the issue explicitly calls out as must-not-happen.
  `LLMAgent._retry_ainvoke`/`_run_pruning_pass`/`stream_message` needed no
  changes; their own outer retry wrapping becomes a harmless no-op for a
  virtual provider since `VirtualChatModel` already exhausts retries
  internally before raising.
- `reasoning_feedback` is read once at the virtual-provider level and
  applied identically to every backend, so a reasoning-aware backend
  switch still round-trips `reasoning_content` correctly.

Also fixes a pre-existing Behave test-isolation bug surfaced while adding
these scenarios: `reasoning_provider_routing_steps.py`'s
`_reasoning_unavailable` context flag wasn't reset in `before_scenario`, so
it silently leaked into any later scenario reusing the shared "the chat
model is constructed" step, breaking unrelated native-provider construction.

Module: src/cleveractors/agents/llm_virtual_provider.py (new),
src/cleveractors/agents/llm_client.py, src/cleveractors/agents/llm.py,
src/cleveractors/validation/_agents.py, src/cleveractors/validation/_vocabulary.py.
BDD: features/virtual_provider.feature. Robot: robot/virtual_provider_routing.robot.
Benchmark: benchmarks/virtual_provider_benchmark.py. Docs: docs/index.md
§4.4/§4.4.7 (Version 1.8.0, §21.1), docs/guides/virtual-provider-backend-routing.md.
Example: examples/packages/virtual-provider-pooled-agent.yaml.

ISSUES CLOSED: #155

Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>

CoreRasurae added a new dependency 2026-10-16 16:51:42 +00:00

#155 Add backends/strategy configuration and round-robin routing for virtual provider LLM agents

CoreRasurae force-pushed feature/m1-virtual-provider-round-robin from 9a16cc5955 to fd8f9629ee 2026-10-16 17:04:56 +00:00 Compare

hurui200320 approved these changes 2026-10-19 14:11:48 +00:00

hurui200320 left a comment

Copy link

PR Review: !157 (Ticket #155)

Verdict: Approve

No critical or major issues found. The retry-pinning design is sound (verified against call_with_retry's control flow: an inner ExecutionError from exhausted retries is re-raised immediately, never re-retried or re-routed by the outer wrapper), round-robin wraps correctly, per-backend credential isolation is real, and the Behave/Robot suites assert actual behavior. The remaining findings are minor documentation/edge-case items — none block merge, but a few are trivial to fold in before it.

Scope note: no review comments, PR comments, or ticket comments exist anywhere, so there are no author decisions to honor. The random strategy is explicitly deferred to issue #156 and was not flagged. Red-tape items (commit footer format, labels) excluded per review instructions.

Critical Issues

None

Major Issues

None

Minor Issues

1. docs/index.md lines 320–324 — duplicated table rows in the normative standard. The §4.4 field table contains the backends and strategy rows twice (four consecutive rows instead of two, exact duplicates). Recommendation: delete the second pair.

2. examples/packages/virtual-provider-pooled-agent.yaml line 10 — example advises an unsupported value. The comment says "set strategy: random for uniform random selection instead", but random is not supported (VALID_VIRTUAL_PROVIDER_STRATEGIES = {"round_robin"}); following the shipped example produces a ConfigurationError. It also contradicts the guide ("round_robin is currently the only supported value"). Recommendation: reword to state that random is planned as a follow-up (issue #156).

3. src/cleveractors/agents/llm_virtual_provider.py lines 365–366 — explicit provider: null/model: null in a backend entry is inconsistent between validation and construction. The validation layer treats None as absent (skips checks), but construction uses raw_backend.get("provider", provider), which returns None when the key exists with a null value — so None.lower() raises AttributeError, wrapped as the confusing Failed to initialize backends[i] (None/gpt-4) instead of inheriting the agent's default. Recommendation: normalize None to the agent default (or reject it explicitly in validation).

4. src/cleveractors/agents/factory.py lines 533–544 — credential-injection-mode gap for virtual agents. In ADR-2026 injection mode, AgentFactory requires credentials[provider] to exist for any llm agent and raises missing credentials for provider — yet that looked-up slice is then entirely ignored by the virtual path (build_chat_model dispatches to build_virtual_chat_model before touching it). A backends-only agent (per-backend credentials, no top-level provider entry) cannot run in injection mode without an otherwise-unused dummy entry. Pre-existing gate, but the interaction is new, untested, and undocumented. Recommendation: skip the provider-credentials requirement when backends is present, or document it.

5. Coverage 96.9% vs the 97% figure (ticket subtask / noxfile.py docstring lines 157–160; actual gate constant is the pre-existing 96.5, so CI is green). The gap is the new sync paths VirtualChatModel._generate/_stream (lines 184–204), which no Behave scenario exercises — all steps use ainvoke/astream. Recommendation: add two cheap scenarios using the sync .invoke()/.stream() (fakes already implement both), or explicitly reconcile the 96.5 vs 97.0 mismatch in the noxfile docstring.

Nits

1. docs/index.md line 511 (§4.4.7) — the body lists random as a strategy without a "not yet supported" marker; only the revision-history row and the guide clarify that. A one-word qualifier would keep the normative body consistent with shipped behavior.
2. src/cleveractors/agents/llm_virtual_provider.py line 43 — first cross-layer import of the private cleveractors.validation._vocabulary from the agents layer. Consider a neutral home for the two constants (or the public cleveractors.validation surface).
3. llm_virtual_provider.py lines 251–267 — on a transient first-chunk failure, the abandoned async generator is never aclose()d before the retry creates a new one. This exactly mirrors the pre-existing LLMAgent.stream_message pattern (llm.py lines 2134–2148), so it is consistent, not a regression.
4. llm_virtual_provider.py lines 384–390 — the non-ConfigurationError wrap omits the DEBUG-level log of the exception type that build_chat_model's sanitize path emits (llm_client.py lines 150–162); a small diagnostics loss.
5. llm.py lines 823–834 — _run_pruning_pass constructs a fresh VirtualChatModel per pruning call, so pruning always lands on backend 0 (fresh round-robin counter each pass). Deterministic and harmless, but an undocumented emergent behavior; pruning timeout messages also lose the provider URL (_get_provider_url finds no URL attrs on the wrapper).
6. features/virtual_provider.feature — no scenario for explicit strategy: round_robin (only the omitted-default case is tested end-to-end).
7. The reasoning round-trip scenario (feature lines 176–180, steps 715–757) validates ReasoningChatModel payloads directly, bypassing VirtualChatModel. The mechanism it proves (reasoning travels in message kwargs + reasoning_feedback applied identically per backend) is what makes the virtual case work, so it is a legitimate proxy — an end-to-end variant would just be stronger.
8. Branch hygiene: the required docs deliverables (§4.4/§4.4.7, version bump) live in a second commit (fd8f962) rather than the atomic feature commit, so neither commit alone fully implements the issue. Linear history with no merge commits, otherwise clean.

Summary

A well-executed, carefully-scoped increment. VirtualChatModel is a genuine BaseChatModel subclass that composes correctly with the existing stack: backend selection happens exactly once per invocation and is pinned through both the ADR-2032 retry loop and the first-chunk-only streaming retry; the outer _retry_ainvoke wrapping degenerates to a no-op exactly as the ticket requires; per-backend construction reuses the unmodified _build_from_credentials path (including SSRF checks and error sanitization); and the temperature-sync in _select() correctly makes the existing _temperature_override flow work through the virtual layer. All ticket acceptance criteria have corresponding, meaningful Behave scenarios, the Robot test drives the real Executor → LLMAgent → VirtualChatModel chain with only the network boundary stubbed, and the accompanying test-isolation fix in features/environment.py is a genuine pre-existing bug fix. The findings above are polish: fix the two documentation defects (trivial), and consider the null-handling and injection-mode items as fast-follows or fold-ins.

## PR Review: !157 (Ticket #155) ### Verdict: Approve No critical or major issues found. The retry-pinning design is sound (verified against `call_with_retry`'s control flow: an inner `ExecutionError` from exhausted retries is re-raised immediately, never re-retried or re-routed by the outer wrapper), round-robin wraps correctly, per-backend credential isolation is real, and the Behave/Robot suites assert actual behavior. The remaining findings are minor documentation/edge-case items — none block merge, but a few are trivial to fold in before it. *Scope note: no review comments, PR comments, or ticket comments exist anywhere, so there are no author decisions to honor. The `random` strategy is explicitly deferred to issue #156 and was **not** flagged. Red-tape items (commit footer format, labels) excluded per review instructions.* ### Critical Issues None ### Major Issues None ### Minor Issues 1. **`docs/index.md` lines 320–324 — duplicated table rows in the normative standard.** The §4.4 field table contains the `backends` and `strategy` rows twice (four consecutive rows instead of two, exact duplicates). Recommendation: delete the second pair. 2. **`examples/packages/virtual-provider-pooled-agent.yaml` line 10 — example advises an unsupported value.** The comment says *"set `strategy: random` for uniform random selection instead"*, but `random` is not supported (`VALID_VIRTUAL_PROVIDER_STRATEGIES = {"round_robin"}`); following the shipped example produces a `ConfigurationError`. It also contradicts the guide ("`round_robin` is currently the only supported value"). Recommendation: reword to state that `random` is planned as a follow-up (issue #156). 3. **`src/cleveractors/agents/llm_virtual_provider.py` lines 365–366 — explicit `provider: null`/`model: null` in a backend entry is inconsistent between validation and construction.** The validation layer treats `None` as absent (skips checks), but construction uses `raw_backend.get("provider", provider)`, which returns `None` when the key exists with a null value — so `None.lower()` raises `AttributeError`, wrapped as the confusing `Failed to initialize backends[i] (None/gpt-4)` instead of inheriting the agent's default. Recommendation: normalize `None` to the agent default (or reject it explicitly in validation). 4. **`src/cleveractors/agents/factory.py` lines 533–544 — credential-injection-mode gap for virtual agents.** In ADR-2026 injection mode, `AgentFactory` requires `credentials[provider]` to exist for any `llm` agent and raises `missing credentials for provider` — yet that looked-up slice is then entirely ignored by the virtual path (`build_chat_model` dispatches to `build_virtual_chat_model` before touching it). A `backends`-only agent (per-backend credentials, no top-level provider entry) cannot run in injection mode without an otherwise-unused dummy entry. Pre-existing gate, but the interaction is new, untested, and undocumented. Recommendation: skip the provider-credentials requirement when `backends` is present, or document it. 5. **Coverage 96.9% vs the 97% figure (ticket subtask / `noxfile.py` docstring lines 157–160; actual gate constant is the pre-existing 96.5, so CI is green).** The gap is the new sync paths `VirtualChatModel._generate`/`_stream` (lines 184–204), which no Behave scenario exercises — all steps use `ainvoke`/`astream`. Recommendation: add two cheap scenarios using the sync `.invoke()`/`.stream()` (fakes already implement both), or explicitly reconcile the 96.5 vs 97.0 mismatch in the noxfile docstring. ### Nits 1. `docs/index.md` line 511 (§4.4.7) — the body lists `random` as a strategy without a "not yet supported" marker; only the revision-history row and the guide clarify that. A one-word qualifier would keep the normative body consistent with shipped behavior. 2. `src/cleveractors/agents/llm_virtual_provider.py` line 43 — first cross-layer import of the private `cleveractors.validation._vocabulary` from the agents layer. Consider a neutral home for the two constants (or the public `cleveractors.validation` surface). 3. `llm_virtual_provider.py` lines 251–267 — on a transient first-chunk failure, the abandoned async generator is never `aclose()`d before the retry creates a new one. This exactly mirrors the pre-existing `LLMAgent.stream_message` pattern (llm.py lines 2134–2148), so it is consistent, not a regression. 4. `llm_virtual_provider.py` lines 384–390 — the non-`ConfigurationError` wrap omits the DEBUG-level log of the exception type that `build_chat_model`'s sanitize path emits (llm_client.py lines 150–162); a small diagnostics loss. 5. `llm.py` lines 823–834 — `_run_pruning_pass` constructs a fresh `VirtualChatModel` per pruning call, so pruning always lands on backend 0 (fresh round-robin counter each pass). Deterministic and harmless, but an undocumented emergent behavior; pruning timeout messages also lose the provider URL (`_get_provider_url` finds no URL attrs on the wrapper). 6. `features/virtual_provider.feature` — no scenario for *explicit* `strategy: round_robin` (only the omitted-default case is tested end-to-end). 7. The reasoning round-trip scenario (feature lines 176–180, steps 715–757) validates `ReasoningChatModel` payloads directly, bypassing `VirtualChatModel`. The mechanism it proves (reasoning travels in message kwargs + `reasoning_feedback` applied identically per backend) is what makes the virtual case work, so it is a legitimate proxy — an end-to-end variant would just be stronger. 8. Branch hygiene: the required docs deliverables (§4.4/§4.4.7, version bump) live in a second commit (`fd8f962`) rather than the atomic feature commit, so neither commit alone fully implements the issue. Linear history with no merge commits, otherwise clean. ### Summary A well-executed, carefully-scoped increment. `VirtualChatModel` is a genuine `BaseChatModel` subclass that composes correctly with the existing stack: backend selection happens exactly once per invocation and is pinned through both the ADR-2032 retry loop and the first-chunk-only streaming retry; the outer `_retry_ainvoke` wrapping degenerates to a no-op exactly as the ticket requires; per-backend construction reuses the unmodified `_build_from_credentials` path (including SSRF checks and error sanitization); and the temperature-sync in `_select()` correctly makes the existing `_temperature_override` flow work through the virtual layer. All ticket acceptance criteria have corresponding, meaningful Behave scenarios, the Robot test drives the real Executor → LLMAgent → VirtualChatModel chain with only the network boundary stubbed, and the accompanying test-isolation fix in `features/environment.py` is a genuine pre-existing bug fix. The findings above are polish: fix the two documentation defects (trivial), and consider the null-handling and injection-mode items as fast-follows or fold-ins.

CoreRasurae changed target branch from master to develop 2026-10-19 20:45:00 +00:00

CoreRasurae force-pushed feature/m1-virtual-provider-round-robin from fd8f9629ee to 9631ee670a 2026-10-19 20:56:35 +00:00 Compare

CoreRasurae merged commit 9631ee670a into develop 2026-10-19 21:15:07 +00:00

CoreRasurae deleted branch feature/m1-virtual-provider-round-robin 2026-10-19 21:15:08 +00:00

CoreRasurae referenced this pull request 2026-10-19 21:39:35 +00:00

feat(agents): add random backend-selection strategy for virtual provider LLM agents #158

Sign in to join this conversation.

Reviewers

No reviewers

hurui200320

Labels

Clear labels

  

auto/blocked-by-deps

PR blocked by an open issue dependency. Operator must close the dep (or remove the dependency link) before the merge driver can act. Auto-cleared by merge_drive when no open deps remain.

  

auto/ci-timeout

Most recent merge cycle hit CI timeout. Driver excludes this PR while last merge_cycle row is < 30 min old; label persists thereafter as visible history.

  

auto/claimed-implementer

Currently being processed by an implementer worker.

  

auto/claimed-merge

Currently being processed by the merge driver.

  

auto/claimed-reviewer

Currently being processed by a reviewer worker.

  

auto/driver-down

Merge driver heartbeat stale; pipeline halted. Closed automatically on next clean tick.

  

auto/invariant-violation

Detected master commit violating the strict merge invariant. Tracked as an issue (not a PR label); kept here for label completeness.

  

auto/last-attempt-tier-0

In-cycle escalation: most recent attempt ran at the Tier 0 slot (`tier-0`). Slot's model defined in .opencode/models/tiers.yaml.

  

auto/last-attempt-tier-1

In-cycle escalation: most recent attempt ran at the Tier 1 slot (`tier-1`). Slot's model defined in .opencode/models/tiers.yaml.

  

auto/last-attempt-tier-2

In-cycle escalation: most recent attempt ran at the Tier 2 slot (`tier-2`). Slot's model defined in .opencode/models/tiers.yaml. Gated behind IMPLEMENTER_ESCALATION_TIER2_ENABLED.

  

auto/last-attempt-tier-min

In-cycle escalation: most recent attempt ran at the Tier -1 slot (`tier-min`). Slot's model defined in .opencode/models/tiers.yaml. Suffix is ``-min`` (not ``--1``) so the Forgejo UI reads naturally.

  

Automation Tracking

Tracking issues used by the AI Automation system for agents to communicate and report.

  

auto/needs-conflict-resolution

Rebase conflict needs LLM conflict-resolver.

  

auto/needs-implementer

Failing CI needs implementer attention.

  

auto/postmortem

Documenting a driver incident or rollback.

  

auto/ready-to-merge

Reviewer has APPROVED this PR and no later REQUEST_CHANGES is outstanding. The merge driver requires this label to even consider a PR for merging. Set by the reviewer worker on APPROVE; cleared on REQUEST_CHANGES.

  

auto/restart-throttled

Train repeatedly lost master-tempo races. Driver excludes via merge_cycle until cooldown elapses; label persists as visible history.

  

auto/revert

Revert PR backing out an invariant violation. Fast-tracked through the merge driver.

  

auto/sentinel

Sentinel PR duplicated from upstream into a personal fork by tools/duplicate_prs_to_fork.py for pipeline testing. Lives only in the fork; the canonical pipeline never sees it.

  

auto/stale-inactivity

No implementer activity for N days. Flagged for human review. Auto-cleared on next push to head branch.

  

auto/unstable

Repeatedly fails on current master (>= 3 ci-fail-on-rebased-sha releases in 12 h). Excluded from driver until human triage.

  

Blocked

A ticket in a blocked state and unable to complete until some other task is completed first.

  

Bounty

$100

A bounty of $100 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$1000

A bounty of $1000 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$10000

A bounty of $10000 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$20

A bounty of $20 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$2000

A bounty of $2000 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$250

A bounty of $250 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$50

A bounty of $50 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$500

A bounty of $500 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$5000

A bounty of $5000 for any open-source contributor who provides a MR that solves this issue

  

Bounty

$750

A bounty of $750 for any open-source contributor who provides a MR that solves this issue

  

MoSCoW

Could have

Could have feature in order to satisfy the epic/legendary.

  

MoSCoW

Must have

Must have feature in order to satisfy the epic/legendary.

  

MoSCoW

Should have

Should have feature in order to satisfy the epic/legendary.

  

Needs Feedback

There are questions in the ticket that can not be completed until the project owner provides clarity.

  

Points

1

1 man-hours worth of work for an expert with no learning curve.

  

Points

13

13 man-hours worth of work for an expert with no learning curve.

  

Points

2

2 man-hours worth of work for an expert with no learning curve.

  

Points

21

21 man-hours worth of work for an expert with no learning curve.

  

Points

3

3 man-hours worth of work for an expert with no learning curve.

  

Points

34

34 man-hours worth of work for an expert with no learning curve.

  

Points

5

5 man-hours worth of work for an expert with no learning curve.

  

Points

55

55 man-hours worth of work for an expert with no learning curve.

  

Points

8

8 man-hours worth of work for an expert with no learning curve.

  

Points

88

88 man-hours worth of work for an expert with no learning curve.

  

Priority

Backlog

This ticket has backlogged priority and is not to be worked on yet

  

Priority

CI Blocker

Critical priority issue that blocks CI/CD pipeline and prevents PR merges

  

Priority

Critical

The priority is critical

  

Priority

High

The priority is high

  

Priority

Low

The priority is low

  

Priority

Medium

The priority is medium

  

Signed-off: Owner

When an epic or legendary is in review it must be signed off by owner, tech lead, and scrum master before being marked as completed.

  

Signed-off: Scrum Master

When an epic or legendary is in review it must be signed off by owner, tech lead, and scrum master before being marked as completed.

  

Signed-off: Tech Lead

When an epic or legendary is in review it must be signed off by owner, tech lead, and scrum master before being marked as completed.

  

Spike

A ticket for learning a tool or technology that is needed to be able to do future planning and design.

  

State

Completed

The ticket has been fully implemented, completed, and merged with the source code. This label should only be applied once a ticket is closed.

  

State

Duplicate

A ticket that represents the same content as an existing ticket.

  

State

In Progress

A ticket that is actively being developed.

  

State

In Review

A ticket that has had some code completed to implement but is waiting to pass peer review and is not yet merged in.

  

State

Paused

This ticket's work started but wasn't finished. It's on hold (likely in a feature branch) and will be resumed later, either due to a blocker or a delay.

  

State

Unverified

All new tickets start in this state. A developer may set it to show the ticket is unverified. This means we haven't agreed to work on it. It will either move to a verified state or be closed as wontdo.

  

State

Verified

The issue has been verified by a developer as legitimate. It will be worked on and verified tickets are now considered part of the backlog.

  

State

Wont Do

This ticket has been decided it wont be done. This may mean the bug has been determined to not be real (cant verify) or the feature is one we have decided we dont want to adopt.

  

Type

Automation

Any edits or discussion about the AI automated coding system.

  

Type

Bug

Something that doesnt work as intended.

  

Type

Discussion

Anytime a ticket represents a discussion about a subject and doesnt fall into one of the other categories.

  

Type

Documentation

An error or improvement needed in the documentation.

  

Type

Epic

Any first tier epic. That is, an epic which contains only issues as children and will not have sub-epics.

  

Type

Feature

Some new functionality not present.

  

Type

Legendary

A type of Epic which will contain other Epics.

  

Type

Refactor

A code change that restructures existing code without changing its external behavior.

  

Type

Support

Someone needs help using the project.

  

Type

Task

A generic task that doesnt fit into the other type categories.

  

Type

Testing

Work exclusively focusing on fixing or expanding testing.

No labels

auto/blocked-by-deps

auto/ci-timeout

auto/claimed-implementer

auto/claimed-merge

auto/claimed-reviewer

auto/driver-down

auto/invariant-violation

auto/last-attempt-tier-0

auto/last-attempt-tier-1

auto/last-attempt-tier-2

auto/last-attempt-tier-min

Automation Tracking

auto/needs-conflict-resolution

auto/needs-implementer

auto/postmortem

auto/ready-to-merge

auto/restart-throttled

auto/revert

auto/sentinel

auto/stale-inactivity

auto/unstable

Blocked

Bounty

$100

Bounty

$1000

Bounty

$10000

Bounty

$20

Bounty

$2000

Bounty

$250

Bounty

$50

Bounty

$500

Bounty

$5000

Bounty

$750

MoSCoW

Could have

MoSCoW

Must have

MoSCoW

Should have

Needs Feedback

Points

1

Points

13

Points

2

Points

21

Points

3

Points

34

Points

5

Points

55

Points

8

Points

88

Priority

Backlog

Priority

CI Blocker

Priority

Critical

Priority

High

Priority

Low

Priority

Medium

Signed-off: Owner

Signed-off: Scrum Master

Signed-off: Tech Lead

Spike

State

Completed

State

Duplicate

State

In Progress

State

In Review

State

Paused

State

Unverified

State

Verified

State

Wont Do

Type

Automation

Type

Bug

Type

Discussion

Type

Documentation

Type

Epic

Type

Feature

Type

Legendary

Type

Refactor

Type

Support

Type

Task

Type

Testing

Milestone

Clear milestone

No items

No milestone

v2.1.0

Projects

Clear projects

No items

No project

Assignees

Clear assignees

No assignees

2 participants

Notifications

Subscribe

Due date

The due date is invalid or out of range. Please use the format "yyyy-mm-dd".

No due date set.

Blocks

#155 Add backends/strategy configuration and round-robin routing for virtual provider LLM agents

cleveragents/cleveractors-core

Reference

cleveragents/cleveractors-core!157

No description provided.