Applies the same cleanup as knowledge-synthesizer to the rest of the
category, all of which shared the identical keyword-stuffed template:
agent-organizer, context-manager, error-coordinator,
multi-agent-coordinator, performance-monitor, task-distributor,
workflow-orchestrator
For each file:
- Removed fabricated metrics — the numeric-target checklists
(">95%", "<100ms", "100%"), the progress/status JSON blobs with
invented counts, and the "Delivery notification" brag sentences the
model would otherwise hallucinate into its output.
- Cut technique/buzzword blocks the tools (Read, Write, Edit, Glob,
Grep) cannot execute; reframed each agent honestly as planning and
documenting rather than running a live runtime.
- Replaced the circular "Communication Protocol" JSON (which assumed a
non-existent message bus) with honest file-and-orchestrator
coordination and explicit tool-scope/grounding rules.
- Preserved frontmatter (name/tools/model) and the sibling-agent
cross-references.
Each file dropped from ~286 lines to ~66-101. No version bump needed
beyond the 1.0.4 already applied for knowledge-synthesizer, since all
changes are within the voltagent-meta category.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Rewrites the knowledge-synthesizer definition to match what its tools
(Read, Write, Edit, Glob, Grep) can actually do, addressing issue #300:
- Removes fabricated metrics ("Pattern accuracy > 85%", "50k+ entities",
"improving performance by 23%") the model would otherwise hallucinate
into its output.
- Drops technique blocks the tools cannot execute (federated/reinforcement
learning, knowledge-graph building, anomaly/causation inference).
- Replaces the circular "Communication Protocol" JSON (which assumed a
non-existent message bus) with an evidence-cited output schema and an
honest note that coordination happens through shared files.
- Adds hard grounding rules: every pattern must cite path:line evidence
and appear in >=2 sources; no invented counts.
Bumps voltagent-meta 1.0.3 -> 1.0.4 (plugin.json + marketplace.json) to
satisfy the version-bump CI gate.
Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Every other agent in this category grants WebFetch/WebSearch for
competitor/industry research; add them here and use them (competitor
landing page research) rather than leaving them unused.
Fills two gaps in the collection: no agent covered email authentication
(SPF/DKIM/DMARC)/deliverability, and none covered conversion-focused
landing page copywriting specifically. Updates category READMEs, root
README, and bumps plugin/marketplace versions per CONTRIBUTING.md.
Switch all 25 subagents that hardcode `model: opus` to `model: inherit`
so they ride the user's session model (including claude-opus-4-8[1m])
instead of silently falling back to standard Opus.
Problem: The `opus` alias always resolves to standard Opus, never the
1M-context variant (claude-opus-4-8[1m]). In Claude Code, these two
variants draw from separate usage allowances — standard Opus has a much
smaller budget. When a user runs their main session on opus[1m] and
spawns one of these subagents, the subagent silently drops to standard
Opus and rapidly drains the scarce standard-Opus allowance.
There is no user-side workaround: aliases can't be remapped globally,
plugin agent files are read-only and overwritten on updates, and
same-named agents in ~/.claude/agents/ don't override plugin agents.
Fix: Using `model: inherit` preserves the original intent ('use the
best model for heavy work') while respecting whichever Opus tier the
user's session is actually running on.
Affected categories:
- 01-core-development: design-bridge, graphql-architect, microservices-architect
- 03-infrastructure: cloud-architect, platform-engineer, security-engineer
- 04-quality-security: ad-security-reviewer, ai-writing-auditor,
architect-reviewer, code-reviewer, compliance-auditor, penetration-tester,
powershell-security-hardening, security-auditor
- 05-data-ai: ai-engineer, llm-architect
- 07-specialized-domains: fintech-engineer, healthcare-admin,
payment-integration, quant-analyst, risk-manager
- 08-business-product: license-engineer
- 09-meta-orchestration: codebase-orchestrator, multi-agent-coordinator,
workflow-orchestrator
plugin.json v1.1.0 references two agent files that don't exist in the
repository:
- cost-accounting-performance-reviewer.md
- performance-roi-translator.md
This causes the plugin to fail loading in Claude Code with:
Path not found: <cache>/voltagent-qa-sec/1.1.0/cost-accounting-performance-reviewer.md (agents)
Path not found: <cache>/voltagent-qa-sec/1.1.0/performance-roi-translator.md (agents)
Drop the orphan entries to restore plugin loading. The two referenced
agents can be re-added later once the corresponding .md files are
committed.
Enhanced the role description for the UI/UX Tester to emphasize the importance of visual spacing and micro-interactions. Updated the testing checklist and methodologies to include specific focus on spacing and granular functionality.
Necati Özmen