Source

• Request id: hydr_9287dc4fd0
• Source path: /tmp/forge_standout_md_pack/13-micro-agents-and-micro-packs.md
• Status: content-seed
• Intended uses: article creation, idea evolution, blog post drafting, architecture documentation hydration
• Tags: micro-agents, local-llm, bounded-judgment, cost, risk
• Body words: 703

Human-Oriented Pattern

• User problem: Readers need a human-oriented explanation of 'Micro-Agents and Micro-Packs' before they can act on the technical detail.
• Outcome: They understand the business meaning, the Forge mechanism, and the next safe step.
• Forge mechanism: Forge Platform connects methodology, contracts, workcells, evidence, and approval boundaries.
• Trust boundary: The hydration agent generates review drafts only; humans choose whether and where to promote them.
• Next action: Review the destination candidates, then promote only the draft that matches source-of-truth ownership.

Destination Candidates

forge_platform_architecture

• Canonical root: /home/lzvyahin/Code/forge-platform/docs
• Draft path: drafts/forge_platform_architecture/micro-agents-and-micro-packs.md
• Audience: Operators and architects integrating Forge Platform components.
• Content type: architecture_note
• Score: 4
• Matched terms: architecture, architecture documentation hydration
• Rationale: The seed explains ecosystem boundaries, platform contracts, or control-plane mechanics.
• CTA candidates:
• Open the platform reference architecture.
• Review the workcell catalog before adding execution code.

• Confirm product boundaries before promoting implementation claims.

• Promotion notes:

• Keep runtime implementation claims aligned with the platform/workcells split.
• If this becomes canonical architecture, add it to the relevant handbook index.

forgesdlc_blog

• Canonical root: /home/lzvyahin/Code/forgesdlc/blog
• Draft path: drafts/forgesdlc_blog/micro-agents-and-micro-packs.md
• Audience: Delivery leaders and engineering leaders evaluating Forge.
• Content type: blog_post
• Score: 4
• Matched terms: article, blog
• Rationale: The seed reads as public ForgeSDLC positioning or operating-model narrative.
• CTA candidates:
• Read the autonomy and bounded execution series.
• Compare ForgeSDLC with generic agent orchestration.

• Use the Forge principles to shape the next delivery conversation.

• Promotion notes:

• Promote only product narrative, not canonical platform contracts.
• Check the blog index and cross-link related ForgeSDLC articles.

forge_lcdl_docs

• Canonical root: /home/lzvyahin/Code/forge-lcdl/docs
• Draft path: drafts/forge_lcdl_docs/micro-agents-and-micro-packs.md
• Audience: Engineers designing governed LLM calls and task contracts.
• Content type: operator_guide
• Score: 2
• Matched terms: llm, task
• Rationale: The seed discusses reliable LLM task design, evaluations, or schema-aware outputs.
• CTA candidates:
• Design the task contract before wiring a workflow.

• Add evals before treating LLM output as automation input.

• Promotion notes:

• Promote only LCDL-specific task and transport guidance.

Overclaim Risks

• Do not describe generated drafts as canonical before human promotion.
• Do not invent customers, certifications, security guarantees, metrics, or integrations.
• Do not imply Forge Platform executes workcells directly; execution belongs to product runners or forge-workcells.
• Do not claim fully autonomous delivery beyond documented autonomy levels and gates.

Seed Excerpt

Micro-Agents and Micro-Packs ## Core thesis Forge's micro-agent idea is innovative because it resists the impulse to turn every task into a large, open-ended agent loop. A micro-agent is a bounded local-LLM run for a single judgment ta...