Academic Pipeline — 10-Stage End-to-End (Hermes Edition)

📄 License: CC BY-NC 4.0 · Copyright (c) 2026 Cheng-I Wu
🔗 Original: Imbad0202/academic-research-skills
🔄 Adaptation: Orchestration pipeline implemented via delegate_task instead of Claude Code's internal agent system. All agent definitions, references, and quality standards preserved unchanged from original. This adaptation is distributed under the same CC BY-NC 4.0 license.

Orchestrates the full academic workflow from research idea to publication-ready output.

10 Stages

Hermes Execution

The orchestrator (main Hermes agent):

1. Stage 1: Clarify user intent → select modes → create Project Plan
2. Stage 2-6: Sequential delegate_task calls, each loading the relevant skill
3. Stage 7-9: Lightweight parallel or sequential based on dependencies
4. Stage 10: Write final package to disk

Key Checkpoints

• Stage 1→2: User approves Project Plan
• Stage 2→3: Research report accepted by user
• Stage 4→5: Review decision received
• Stage 6→7: Polished draft ready

Resume Support

Pipeline state stored in ~/.hermes/sessions/. User can resume with:

Continue the academic pipeline from Stage 4

Usage Patterns

Full pipeline:

Take "AI impact on HE quality assurance" through the full pipeline to publication-ready manuscript

Partial pipeline:

Start from Stage 3 with my existing deep-research output

Resume:

Resume pipeline from Stage 5 with reviewer comments: [comments]

Integration with Hindsight Memory

Pipeline state and key outputs are retained to Hindsight for long-term tracking:

hindsight memory retain --content "[stage_output]" --tags "pipeline,stage-N"

Critical Rules

1. ⚠️ Each stage MUST complete before the next begins
2. ⚠️ User confirmation at all IRON RULE checkpoints
3. ⚠️ Ground truth isolation: review stage must NOT see author identity

Security & Privacy

Orchestration disclosure: This pipeline orchestrates multiple skills (research → writing → review) via delegate_task. All content passes through the AI provider's workflow. Use only with topics you are comfortable processing through AI.

Tool access: This pipeline delegates to sub-skills which independently control their own tool access. No stage is granted terminal, web, or system tools beyond what each sub-skill permits.

Agent files: All agents/ directories across sub-skills contain academic prompt templates, NOT system prompt overrides.