chore: enable emoji reactions for agent runs on this repo

[ralphbean]

Enable emoji reactions for this repo.

[qodo-code-review]

PR Summary by Qodo

Enable emoji reactions for agent run status

⚙️ Configuration changes 🕐 Less than 5 minutes

AI Description

• Enables emoji reactions when agent runs start.
• Enables emoji reactions when agent runs complete.

High-Level Assessment

The repository-level status notification configuration is the appropriate approach because it uses the existing Fullsend feature directly; custom automation or code changes would add unnecessary complexity.

Files changed (1) +4 / -0

Other (1) +4 / -0

config.yamlEnable agent-run reaction notifications +4/-0

Enable agent-run reaction notifications

• Adds repository-level Fullsend status notification settings that enable emoji reactions at agent-run start and completion.

.fullsend/config.yaml

[fullsend-ai-review]

🤖 Finished Review · ✅ Success · Started 2:20 PM UTC · Completed 2:27 PM UTC

Commit: 6127703 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $1.32

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[qodo-code-review]

Code Review by Qodo

🐞 Bugs (0) 📘 Rule violations (0) 📎 Requirement gaps (0)

Great, no issues found!

Qodo reviewed your code and found no material issues that require review

Tip of the day

💡 Did you know, you can hide the parts of a finding you never read, like the evidence or the agent prompt

More tips ↗ | Customize Qodo ↗ | Qodo docs ↗

[fullsend-ai-review]

Looks good to me

[waynesun09]

Approving at 6127703.

Four-line addition of status_notifications.reaction.{start,completion} to .fullsend/config.yaml. Verified the key is real and wired rather than a silent no-op — status_notifications is consumed in internal/config/config.go, internal/harness/forge.go and internal/cli/reconcilestatus.go, with coverage in internal/cli/reconcilestatus_test.go, internal/config/config_test.go and pkg/behaviourtest/steps/reaction.go, and it is documented in the layered-config reference. That check matters here because fullsend run does not call Validate(), so an unsupported key would fail silently.

Config-only, no code paths touched, CI green, no unresolved review threads.

[fullsend-ai-retro]

🤖 Finished Retro · ✅ Success · Started 7:52 PM UTC · Completed 8:04 PM UTC

Commit: 6127703 · View workflow run →

Runtime: claude · Model: opus → claude-opus-4-6 · Effort: high · Cost: $3.69

[fullsend-ai-retro]

Retro: PR #6594 — enable emoji reactions config

PR: #6594 — human-authored 4-line config change adding status_notifications.reaction.{start,completion}: enabled to .fullsend/config.yaml.

Timeline

1. 14:18 UTC — PR opened by ralphbean.
2. 14:18–14:27 UTC — Review agent ran (workflow 32858656116). Classified the change as "mechanical / value-only (feature flag toggle)" with a trivial scope constraint (≤5 tool calls, no surrounding-file reads). Dispatched 3 sub-agents (correctness, security, style-conventions). All returned zero findings. Approved with "Looks good to me."
3. 18:11 UTC — Human reviewer waynesun09 approved with a detailed rationale: verified the config keys are real and wired into the codebase by tracing consumption through internal/config/config.go, internal/harness/forge.go, and internal/cli/reconcilestatus.go, with test coverage in three test files. Explicitly noted this check mattered because fullsend run does not call Validate(), so unsupported keys would fail silently.
4. 19:49 UTC — PR merged.

Key finding

The review agent's "trivial" scope constraint prevented its correctness sub-agent from performing consumer-completeness verification — the exact check the human reviewer identified as critical. The orchestrator misclassified the change as "value-only (feature flag toggle)" when it actually introduces a new top-level config key namespace (status_notifications). This is structurally different from toggling an existing flag: new keys require verifying a code-side consumer exists, otherwise they fail silently.

Additionally, the orchestrator's own reasoning detected a genuine indentation inconsistency (the existing file uses 4-space indentation; the new block uses 2-space), but the orchestrator was constrained to only merge sub-agent findings and could not report its own. The style-conventions sub-agent had already exited early under its "value-only change" criteria and incorrectly claimed the formatting matched the existing file — it only had the diff, not the full file, so it couldn't verify the claim.

Corroborating evidence for existing issues

• #5777 (Review model routing should use diff complexity): This review cost $1.32 (Opus) for a 4-line config change. A cheaper model with appropriate consumer-completeness tooling would have been more cost-effective.
• #2836 (Add scope_constraint handling to sub-agents): This retro confirms the correctness sub-agent lacks scope_constraint-aware early-exit criteria. However, Add scope_constraint handling and exploration budgets to correctness, docs-currency, and cross-repo-contracts sub-agents #2836 addresses propagating constraint enforcement — it does not address the upstream classification accuracy problem found here (misclassifying structural config changes as value-only).
• #4060 (Reduce sub-agent dispatch for mechanical changes): Complementary — addresses dispatch breadth for trivial PRs. This retro's finding is about classification accuracy, which is upstream of both pr-review skill: reduce sub-agent dispatch for mechanical/value-only changes like dependency bumps #4060's breadth optimization and Add scope_constraint handling and exploration budgets to correctness, docs-currency, and cross-repo-contracts sub-agents #2836's depth enforcement.

Proposals filed

• Review orchestrator should not classify config changes introducing new keys as value-only (in fullsend-ai/agents)