2025-08-30 - 2026-08-30
Overview
63 Pull requests merged by 5 users
Merged
#122 fix(agents): wire XAI_API_KEY through reusable workflow reliably
Merged
#121 feat(agents): automatic Ollama Cloud → xAI fallback
Merged
#120 Allow hermes user to trigger AI agent workflow
Merged
#119 fix(agents): rename xai provider key to xai-oc to avoid opencode catalog collision
Merged
#115 hotfix: wire the xai provider into install-opencode.sh
Merged
#114 chore: switch @lead/@ops from Anthropic to xAI grok-4.5
Merged
#113 @junior: publish.sh: PR-creation and push failures are silently swallowed
Merged
#109 fix: deterministic dedup of agent double-posts (self-post + framework reply)
Merged
#108 feat(ci): re-enable dependency caching (runner cache server fixed)
Merged
#107 revert(ci): remove cache steps until runner cache server works
Merged
#106 feat(ci): dependency caching for agent runs (opencode + Playwright)
Merged
#105 fix(agent): reactions show the agent, not "Ghost"
Merged
#104 fix(run-agent): stop agents self-posting their reply as a duplicate comment
Merged
#102 agents: discussion mode — devs talk in threads, build only on explicit signal
Merged
#96 agents: review round 2 — split-PR flow, thread recency, mention boundaries, learnings cap, pm model
Merged
#95 agents: hard timeout on agent runs (hung model blocked the runner ~1h)
Merged
#93 @senior: model name in comment
Merged
#91 gitea-admin skill: packages / container-registry section
Merged
#90 report: static details label on comment dropdowns
Merged
#89 report: one uniform dropdown (tokens + tool calls) on every agent comment
Merged
#87 @senior: retro: issue #83
Merged
#88 publish: qa approval detail on the PR, terse verdict on the issue
Merged
#84 @junior: intern
Merged
#85 agents: retro loop — @pm retro distills learnings into LEARNINGS.md
Merged
#81 publish: trigger comments speak in first person
Merged
#77 agents: review fixes — thread attribution, concurrency, rescue stall, docs, CI, pin, cost label
Merged
#75 agents: report tokens + $ cost on every agent comment
Merged
#74 chore: revert PM-flow test edit
Merged
#73 @junior: test: PM flow autopilot
Merged
#69 agents: PM-orchestrated flow (@pm hub, @qa reviews, @pm/human merges)
Merged
#68 agents: fix ${SECRETS_JSON:-{}} brace bug (caller skills silently skipped)
Merged
#67 agents: document single-line requirement for caller-skill secrets
Merged
#66 agents: drop redundant agent-name headers from comments (+ fix self-trigger loop)
Merged
#65 agents: caller-provided skills hook (move node1-ssh to the owning repo)
Merged
#63 agents: remove shared AGENT_TOKEN — per-agent tokens only
Merged
#62 docs: README reflects per-agent SELF_TOKEN model
Merged
#59 agents: per-agent Gitea identity — each agent uses its own token
Merged
#58 feat(ops): @ops uses dedicated TOKEN_OPS not AGENT_TOKEN
Merged
#57 ops: .env is the token source of truth (not tokens.md)
Merged
#56 feat(ops): make @ops routable + TOKEN_OPS identity
Merged
#55 feat(ops): add @ops Gitea-administrator agent
Merged
#52 fix(autopilot): merge via AGENT_TOKEN so it triggers deploy
Merged
#51 autopilot(qa): bounce to dev instead of halt; stop after 3
Merged
#20 @lead: run without asking user
Merged
#50 fix(activity-log): tool calls only (drop prose summary)
Merged
#49 fix(activity-log): strip PR-description block from reasoning trail
Merged
#47 fix(route): resume existing ai/issue-N branch on re-trigger
Merged
#45 fix: quote run-name (bare # was a YAML comment)
Merged
#43 chore: title agent runs by issue number (run-name)
Merged
#41 chore: standardize the ai-agent caller stub
Merged
#40 fix(routing): expose agent skills in the roster
Merged
#39 @senior: One comment per agent run
Merged
#36 @junior: Add runtime .agents-workflow/ checkout to .gitignore
Merged
#32 Single source of truth for agent model IDs (fixes #31)
Merged
#34 @lead: Dev-agent runs can break later steps by touching the runtime .agents-workflow scripts checkout
Merged
#25 @lead: agent skills
Merged
#24 @junior: run on issue
Merged
#15 @senior: gitea
Merged
#13 @junior: agents ssh
Merged
#11 @senior: Wire up ssh node1 … in the agent workflow (read-only swarm diagnostics)
Merged
#9 @senior: readme
Merged
#5 @senior: opscode events are not in the PR
Merged
#3 @senior: opencode logs
1 Pull request proposed by 1 user
Proposed
#117 @junior: lead update to kimi-k3
37 Issues closed from 4 users
Closed
#118 @ops/@lead (xai/grok-4.5) crash on every run: opencode provider-name collision with built-in xai catalog entry
Closed
#111 PR-creation step can silently fail, leaving an orphan branch with no PR
Closed
#112 publish.sh: PR-creation and push failures are silently swallowed
Closed
#101 Model interview
Closed
#103 test: discussion mode
Closed
#100 test: intern in gitea/agents
Closed
#97 test: pm investigation
Closed
#98 test: split-PR resolve
Closed
#92 model name in comment
Closed
#86 retro: issue #83
Closed
#83 intern
Closed
#82 intern
Closed
#53 Intern
Closed
#78 test: review-fixes smoke
Closed
#76 test: cost footer
Closed
#72 test: PM flow autopilot
Closed
#70 test: PM-orchestrated flow (normal)
Closed
#64 test: delegation trigger without AGENT_TOKEN
Closed
#60 test: per-agent identity smoke test
Closed
#16 run without asking user
Closed
#44 runname check A
Closed
#46 runname verify A
Closed
#42 RUNNAME-TEST agents
Closed
#38 One comment per agent run
Closed
#35 Add runtime .agents-workflow/ checkout to .gitignore
Closed
#31 Prevent agent model-ID drift between route.sh and install-opencode.sh
Closed
#33 Dev-agent runs can break later steps by touching the runtime .agents-workflow scripts checkout
Closed
#29 De-duplicate agent model IDs (registry vs provider-config drift)
Closed
#28 test routing fix
Closed
#22 agent skills
Closed
#19 auto start pm on issue
Closed
#23 run on issue
Closed
#14 gitea
Closed
#12 agents ssh
Closed
#7 new agent called intern
Closed
#10 Wire up ssh node1 … in the agent workflow (read-only swarm diagnostics)
Closed
#8 readme
39 Issues created by 1 user
Opened
#7 new agent called intern
Opened
#8 readme
Opened
#10 Wire up ssh node1 … in the agent workflow (read-only swarm diagnostics)
Opened
#12 agents ssh
Opened
#14 gitea
Opened
#16 run without asking user
Opened
#19 auto start pm on issue
Opened
#22 agent skills
Opened
#23 run on issue
Opened
#28 test routing fix
Opened
#29 De-duplicate agent model IDs (registry vs provider-config drift)
Opened
#31 Prevent agent model-ID drift between route.sh and install-opencode.sh
Opened
#33 Dev-agent runs can break later steps by touching the runtime .agents-workflow scripts checkout
Opened
#35 Add runtime .agents-workflow/ checkout to .gitignore
Opened
#38 One comment per agent run
Opened
#42 RUNNAME-TEST agents
Opened
#44 runname check A
Opened
#46 runname verify A
Opened
#53 Intern
Opened
#60 test: per-agent identity smoke test
Opened
#64 test: delegation trigger without AGENT_TOKEN
Opened
#70 test: PM-orchestrated flow (normal)
Opened
#72 test: PM flow autopilot
Opened
#76 test: cost footer
Opened
#78 test: review-fixes smoke
Opened
#82 intern
Opened
#83 intern
Opened
#86 retro: issue #83
Opened
#92 model name in comment
Opened
#97 test: pm investigation
Opened
#98 test: split-PR resolve
Opened
#100 test: intern in gitea/agents
Opened
#101 Model interview
Opened
#103 test: discussion mode
Opened
#111 PR-creation step can silently fail, leaving an orphan branch with no PR
Opened
#112 publish.sh: PR-creation and push failures are silently swallowed
Opened
#116 lead update to kimi-k3
Opened
#118 @ops/@lead (xai/grok-4.5) crash on every run: opencode provider-name collision with built-in xai catalog entry
Opened
#123 run-agent.sh: prompt passed as a single argv exceeds MAX_ARG_STRLEN on long threads