desplega-ai/agent-swarm v1.129.0

on GitHub

Install & Run

npm:

npx @desplega.ai/agent-swarm@1.129.0
# or install globally
npm i -g @desplega.ai/agent-swarm@1.129.0

bun:

bunx @desplega.ai/agent-swarm@1.129.0
# or install globally
bun i -g @desplega.ai/agent-swarm@1.129.0

Docker:

# Server
docker pull ghcr.io/desplega-ai/agent-swarm:1.129.0
# Worker
docker pull ghcr.io/desplega-ai/agent-swarm-worker:1.129.0

E2B templates:

# API
agent-swarm-api-1-129-0
agent-swarm-api-latest

# Worker / lead runtime
agent-swarm-worker-1-129-0
agent-swarm-worker-latest

What's Changed

• fix(slack): finalize accepted-message reactions by @desplega-bot in #1103
• docs(thoughts): brainstorm + research — real-time collab primitive (CRDT rooms on KV) for #1090 by @tarasyarema in #1104
• Skills to DB seeding (Phases 1-3): single-source SKILL.md twins, migrate plugin/skills, vendor ai-toolbox by @tarasyarema in #1083
• feat(workflows): foreach node — fan-out agent-task children with parent join by @tarasyarema in #1093
• fix(entrypoint): env-pin git author and committer identity by @fuvidani in #1105
• feat(apps): Swarm Apps — versioned lifecycle, schema-change engine, reusable elements, RBAC, per-user config by @tarasyarema in #1066
• chore(skills): finalize generic seeded templates by @desplega-bot in #1106
• fix: block SSRF through MCP server proxy by @desplega-bot in #1107
• chore(deps): bump context-mode to 1.0.169 by @desplega-bot in #1108
• chore(deps): bump @desplega.ai/agent-fs to 0.12.2 by @desplega-bot in #1109
• fix(ui): unblock Save on blessed connections and stop clobbering OpenAPI base URLs by @tarasyarema in #1110
• fix(slack): unify persona and steering reactions by @desplega-bot in #1111
• fix: preserve profile edits before boot overwrite by @desplega-bot in #1116
• feat(usage): per-user cost + attribution coverage on the Usage page by @desplega-bot in #1115
• Reject incomplete required approval responses by @desplega-bot in #1119
• fix(attachments): encode download filenames and link agent-fs rows by @fuvidani in #1118
• fix(slack): finalize render-v2 reactions by @desplega-bot in #1121
• fix(helm): keep agent-fs backend image current (0.9.0 → 0.12.2) + GHCR drift check by @desplega-bot in #1127
• fix(ui): treat superseded tasks as terminal by @desplega-bot in #1128
• fix: preserve attachment scope and filename integrity by @desplega-bot in #1126
• fix(internal-ai): harden structured summaries and unblock pre-push by @desplega-bot in #1125
• feat(scripts): per-app generated TS types for swarm scripts (DES-767) by @tarasyarema in #1130
• fix(harness): codex boot-gate/pool credential mismatch + MODEL_OVERRIDE reload clobber by @desplega-bot in #1114
• fix(github): require bot identity for review replies by @desplega-bot in #1133
• fix(attribution): repair requester origins and history by @desplega-bot in #1132
• feat(mcp): publish host-agnostic registry entry by @desplega-bot in #1135
• fix(security): harden page-session secret, page-proxy allowlist, page ownership (Superagent cluster A) by @desplega-bot in #1113
• fix: restore session summary transcript sourcing by @desplega-bot in #1131
• fix(costs): TTL-aware, per-model session cost recompute with harness dual-write by @tarasyarema in #1134
• fix(security): sandbox script execution + fail-closed webhook auth (superagent c27edfd7) by @desplega-bot in #1112
• fix: route session summaries through compiled CLI by @desplega-bot in #1137
• fix(claude-adapter): drop malformed modelUsage breakdowns instead of zero-filling by @tarasyarema in #1136
• fix(security): stop bun -e from re-parsing argv-injected flags as code (P1) by @desplega-bot in #1138
• feat(ui): theme presets + per-app theming, motion layer, hive refinements, json-render quality slice 1 by @tarasyarema in #1123
• release: v1.129.0 + docs daily update 2026-08-07 by @desplega-bot in #1139

Full Changelog: v1.128.0...v1.129.0