Skip to content
Uvilo OS
approved Visibility internal Owner erik@uvilo.com Approver _ Created _ Updated _

Knowledge Consistency Plan 7

Plan 7 — Bot Prompts, Available Skills, and Follow-ups

Scope: Update Bot system prompts (database records): align Project Runner default gates and remove Pending state, fix Project Worker prompt typo and publish it, fix Page Manager prompt tool name, update Project Evaluator and Project Worker prompts for Page Manager routing, update the Available Skills list in the Persistent Knowledge prompt, remove MCP references from documentation, ensure run/cost evidence in handoff packets, and create follow-up todo items for platform code changes.

Spec sections covered: S1.9, S1.10 (bot prompts only — Project Evaluator and Project Worker), S3.1 (Project Runner prompt only), S3.2, S3.3, S3.4, S5, S9.1, S9.2, S11.1, S11.2, S12.1 (Project Runner prompt only), S12.2, S12.3 (Project Runner prompt only).

Prerequisite: Plan 2 (skill rename Orchestration to Agentic_Prompts must complete before updating the Available Skills list in Task 5).

Task 1 — Update Project Runner Bot Prompt

Spec S3.1: “The default human gates are consistent across all sources. Project Runner bot prompt: Vision_Eval, Spec_Eval, Plan_Eval, Verify.” Spec S12.1: “The valid Phase: value statuses are Started, Completed, and Blocked only. There is no Pending state.” Spec S12.2: “Every project skill/agent is responsible for setting its own phase to _Started at the beginning of its work and to _Completed once the work is done. The Project Runner does not pre-set a phase to a ‘Pending’ value awaiting agent start; it dispatches the responsible agent, which sets _Started itself on beginning and _Completed on finishing.” Spec S12.3: “Project Runner (prm_cmq8aja9i000301mzab1ezdxf, handle project-runner): Phase Naming Convention: ‘status is Pending, Started, Completed, or Blocked’ → ‘status is Started, Completed, or Blocked’. Autonomous Mode procedure: remove the _Pending handling branch. Routing step: ‘set that next phase to _Started or _Pending as appropriate’ → the next phase is left unset; the Runner spawns the responsible agent, which sets _Started itself.”

Steps

  1. Load the Project Runner bot prompt (prm_cmq8aja9i000301mzab1ezdxf, handle project-runner) using the prompts tool group.
  2. S3.1: Ensure the default human gates listed in the prompt are Vision_Eval, Spec_Eval, Plan_Eval, Verify. Add Vision_Eval if missing.
  3. S12.1: Update the Phase Naming Convention: change “status is Pending, Started, Completed, or Blocked” to “status is Started, Completed, or Blocked”.
  4. S12.2/S12.3: Remove the _Pending handling branch from the Autonomous Mode procedure (“If the phase is _Pending, set it to _Started only for the phase being dispatched, then spawn the responsible agent”). The Runner dispatches the responsible agent, which sets its own phase to _Started.
  5. S12.3: Update the Routing step: change “set that next phase to _Started or _Pending as appropriate” to “the next phase is left unset; the Runner spawns the responsible agent, which sets _Started itself.”
  6. Save the updated prompt using the prompts tool group.

Verify

  • The Project Runner prompt lists default gates as Vision_Eval, Spec_Eval, Plan_Eval, Verify
  • The Phase Naming Convention states only Started, Completed, or Blocked (no Pending)
  • The Autonomous Mode procedure has no _Pending handling branch
  • The Routing step does not reference _Pending

Task 2 — Update Project Worker Bot Prompt

Spec S3.2: “The Project Worker bot prompt’s Standard Handoff Packet table reads Work-Product Path or Link (matching all other bot prompts), not Work-Product Link (double space).” Spec S3.4: “The Project Worker bot prompt (prm_cmqmmf7vr000b01p15biyasvf) has publishedAt set to a timestamp — it is no longer a draft.” Spec S1.10: “Project Worker (prm_cmqmmf7vr000b01p15biyasvf, handle project-worker) — Key Rules: currently directs agents to ‘use the manage-pages skill’ for page creation, moving, deletion, frontmatter updates, and sidebar updates; must be updated to spawn the Page Manager bot via spawnAgent with { botGroup: 'forge', botHandle: 'page-manager' }.”

Steps

  1. Load the Project Worker bot prompt (prm_cmqmmf7vr000b01p15biyasvf, handle project-worker) using the prompts tool group.
  2. S3.2: Fix the Standard Handoff Packet table: change Work-Product Link (double space) to Work-Product Path or Link (matching all other bot prompts).
  3. S1.10: Update the Key Rules section: replace instructions to “use the manage-pages skill” for page creation, moving, deletion, frontmatter updates, and sidebar updates with instructions to spawn the Page Manager bot via spawnAgent with { botGroup: 'forge', botHandle: 'page-manager' }.
  4. S3.4: Set publishedAt to a current timestamp to publish the prompt (it is no longer a draft).
  5. Save the updated prompt using the prompts tool group.

Verify

  • The Standard Handoff Packet table reads Work-Product Path or Link (no double space)
  • The Key Rules section instructs spawning the Page Manager bot, not using the manage-pages skill
  • publishedAt is set to a timestamp (prompt is published)

Task 3 — Update Project Evaluator Bot Prompt

Spec S1.10: “Project Evaluator (prm_cmqmkjxc6000u01o6ddc7p9kn, handle project-evaluator) — Key Rules: currently directs agents to ‘use the manage-pages skill’ for page creation, moving, deletion, frontmatter updates, and sidebar updates; must be updated to spawn the Page Manager bot via spawnAgent with { botGroup: 'forge', botHandle: 'page-manager' }.”

Steps

  1. Load the Project Evaluator bot prompt (prm_cmqmkjxc6000u01o6ddc7p9kn, handle project-evaluator) using the prompts tool group.
  2. S1.10: Update the Key Rules section: replace instructions to “use the manage-pages skill” for page creation, moving, deletion, frontmatter updates, and sidebar updates with instructions to spawn the Page Manager bot via spawnAgent with { botGroup: 'forge', botHandle: 'page-manager' }.
  3. Save the updated prompt using the prompts tool group.

Verify

  • The Key Rules section instructs spawning the Page Manager bot, not using the manage-pages skill
  • No “manage-pages skill” references remain in the prompt

Task 4 — Fix Page Manager Bot Prompt Tool Name

Spec S3.3: “The Page Manager bot prompt references forge-filesystem (the correct MCP server name), not uvilo-filesystem.”

Steps

  1. Load the Page Manager bot prompt (prm_cmqmgkp1v000o01o6kc6b8t6m, handle page-manager) using the prompts tool group.
  2. S3.3: Replace any references to uvilo-filesystem with forge-filesystem (the correct MCP server name).
  3. Save the updated prompt using the prompts tool group.

Verify

  • The Page Manager prompt references forge-filesystem, not uvilo-filesystem

Task 5 — Update Available Skills List in Persistent Knowledge Prompt

Spec S5.1: “The following skills are added to the Available Skills list: artifacts-markdown (Artifacts_Markdown), audit-context (Audit_Context), start-astro-dev (Start_Astro_Dev), test-tools (Test_Tools).” Spec S5.2: “The following skills are removed from the Available Skills list: orchestration — renamed to agentic-prompts (see S1.1); mongodb — does not exist (see S1.8).” Spec S5.3: “Skills Renamed: orchestration to agentic-prompts (see S1.1).” Spec S5.4: “The Available Skills list includes all published skills with folders in Forge/Skills/, excluding Linear_MCP, Suprsend_MCP, and Vercel_MCP (retained on disk but not listed — see S1.9).”

Steps

  1. Load the Persistent Knowledge prompt (prm_cmq83857c000101mzrhmopc03) using the prompts tool group. This prompt contains the Available Skills list.
  2. S5.1: Add the following skills to the Available Skills list: artifacts-markdown (Artifacts_Markdown), audit-context (Audit_Context), start-astro-dev (Start_Astro_Dev), test-tools (Test_Tools).
  3. S5.2: Remove orchestration (renamed to agentic-prompts) and mongodb (does not exist) from the list.
  4. S5.3: Ensure agentic-prompts appears in the list (renamed from orchestration).
  5. S5.4: Ensure the list includes all published skills with folders in Forge/Skills/, excluding Linear_MCP, Suprsend_MCP, and Vercel_MCP (retained on disk but not listed per S1.9).
  6. Save the updated prompt using the prompts tool group.

Verify

  • The Available Skills list includes artifacts-markdown, audit-context, start-astro-dev, test-tools
  • The list does not include orchestration (renamed) or mongodb (removed)
  • The list includes agentic-prompts (renamed from orchestration)
  • The list does not include Linear_MCP, Suprsend_MCP, or Vercel_MCP

Task 6 — Remove MCP References from Documentation

Spec S1.9: “The Linear_MCP, Suprsend_MCP, and Vercel_MCP skills remain on disk but are not listed in the Available Skills list. References to these MCPs are removed from documentation. The skills are retained because they may contain repurposable knowledge for future MCP connections via Composio. Their frontmatter status remains draft.”

Steps

  1. Search all documentation files (.md files in Forge/ root and subdirectories, excluding Forge/Skills/Linear_MCP/, Forge/Skills/Suprsend_MCP/, Forge/Skills/Vercel_MCP/ skill folders themselves) for references to Linear_MCP, Suprsend_MCP, and Vercel_MCP.
  2. Remove references to these three MCP skills from documentation files. This does NOT modify the skill files themselves — only documentation that references them as available/active skills.
  3. Do NOT remove references from: the three MCP skill folders’ own SKILL.md files, the Available Skills list (handled in Task 5), or the project documents (Spec, Plan, Research, Vision, Eval files).
  4. If a documentation file lists these as available skills or recommends their use, remove or update that reference.

Verify

  • Documentation files (excluding the three MCP skill folders and project documents) do not reference Linear_MCP, Suprsend_MCP, or Vercel_MCP as available skills
  • The three MCP skill folders remain on disk with status: draft frontmatter

Task 7 — Ensure Run/Cost Evidence in Handoff Packets and Runs Document

Spec S9.1: “All bot handoff packets include Run/Cost Evidence. When exact evidence is available, it links to the run summary. When exact evidence is unavailable, the packet states ‘exact evidence is unavailable’ as a consistent fallback.” Spec S9.2: “The Runs document ({Project}_Runs.md) is generated during the Verify phase even when exact data is unavailable, recording available context.”

Steps

  1. S9.1: Review all forge-group bot prompts that produce handoff packets (Project Runner, Project Thinker, Project Evaluator, Project Worker, Page Manager, Task Runner). Ensure each prompt’s Standard Handoff Packet table includes a Run / Cost Evidence field with the guidance: “Links to the run summary or records that exact evidence is unavailable.”
  2. For any bot prompt missing the Run/Cost Evidence field, add it to the Standard Handoff Packet table.
  3. S9.2: Ensure the Project_Verify skill (Forge/Forge_Project/Skills/Project_Verify/SKILL.md) or the Project Runner prompt instructs generating {Project}_Runs.md during the Verify phase, even when exact data is unavailable, recording available context.
  4. Save any updated prompts using the prompts tool group.

Verify

  • All forge-group bot prompts with handoff packets include a Run / Cost Evidence field
  • The Project_Verify skill or Project Runner prompt instructs generating {Project}_Runs.md during Verify

Task 8 — Create Follow-up Todo Items for Platform Code Changes

Spec S11.1: “Modify the application/vnd.git-changes artifact type to accept two commit hashes as parameters, displaying the diff between them. This makes git-changes useful for comparing arbitrary commits, not just changes vs upstream.” Spec S11.2: “Update the conversation tools to include usage fields (token counts, model used, cost data) so that Run/Cost evidence can be automatically captured in handoff packets.”

Steps

  1. S11.1: Create a user-owned todo item using createTodoItem with the following body content:
    • Department: Forge
    • Project: Forge/Knowledge_Consistency
    • Phase: Plan (follow-up)
    • Requesting Bot/Role: Project Thinker
    • Action: Modify the application/vnd.git-changes artifact type to accept two commit hashes as parameters, displaying the diff between them.
    • Rationale: Makes git-changes useful for comparing arbitrary commits, not just changes vs upstream. Identified in Spec S11.1.
    • Status: Open
    • Work-Product Link: Forge/Projects/Knowledge_Consistency/Knowledge_Consistency_Spec.md (Section 11.1)
  2. S11.2: Create a user-owned todo item using createTodoItem with the following body content:
    • Department: Forge
    • Project: Forge/Knowledge_Consistency
    • Phase: Plan (follow-up)
    • Requesting Bot/Role: Project Thinker
    • Action: Update the conversation tools to include usage fields (token counts, model used, cost data) so that Run/Cost evidence can be automatically captured in handoff packets.
    • Rationale: Enables automatic capture of Run/Cost evidence in handoff packets. Identified in Spec S11.2.
    • Status: Open
    • Work-Product Link: Forge/Projects/Knowledge_Consistency/Knowledge_Consistency_Spec.md (Section 11.2)

Verify

  • Two todo items exist in the system: one for S11.1 (git-changes artifact enhancement) and one for S11.2 (convo tools usage fields)
  • Each todo item is user-owned and references the Spec section

Task 9 — Build, Commit, and Push

Spec S1.9, S1.10, S3.1–S3.4, S5, S9.1, S9.2, S11.1, S11.2, S12.1–S12.3: This Plan implements bot prompt, Available Skills, and follow-up corrections. After all tasks are complete, commit and push any file changes. Per Known Infrastructure Context, npm run build is expected to fail; proceed with git operations regardless.

Steps

  1. Run npm run build from the repo root. This is known to fail. Proceed regardless.
  2. git add any modified files (e.g., Forge/Forge_Project/Skills/Project_Verify/SKILL.md if modified, documentation files with MCP references removed).
  3. Note: Bot prompt changes are database records, not file changes. They are saved via the prompts tool group, not via git.
  4. git commit -m "Plan 7: Bot prompts, Available Skills, and follow-ups (S1.9, S1.10, S3.1-S3.4, S5, S9.1, S9.2, S11.1, S11.2, S12.1-S12.3)"
  5. git push origin dev

Verify

  • All modified files are committed and pushed to dev
  • git log --oneline -1 shows the commit

Plan 7 Completion Criteria

  • Project Runner prompt lists default gates as Vision_Eval, Spec_Eval, Plan_Eval, Verify; Phase Naming Convention has no Pending; Autonomous Mode has no _Pending handling branch; Routing step does not reference _Pending
  • Project Worker prompt has Work-Product Path or Link (no double space), instructs Page Manager spawn pattern, and is published (publishedAt set)
  • Project Evaluator prompt instructs Page Manager spawn pattern, not manage-pages skill
  • Page Manager prompt references forge-filesystem, not uvilo-filesystem
  • Available Skills list includes artifacts-markdown, audit-context, start-astro-dev, test-tools, agentic-prompts; does not include orchestration, mongodb, Linear_MCP, Suprsend_MCP, Vercel_MCP
  • Documentation files do not reference Linear_MCP, Suprsend_MCP, or Vercel_MCP as available skills
  • All forge-group bot prompts with handoff packets include Run/Cost Evidence field
  • Project_Verify skill or Project Runner prompt instructs generating {Project}_Runs.md during Verify
  • Two follow-up todo items created (S11.1 git-changes enhancement, S11.2 convo tools usage fields)
  • Changes committed and pushed to dev

Dependencies

  • Requires: Plan 2 (skill rename Orchestration to Agentic_Prompts must complete before updating Available Skills list in Task 5)
  • Blocks: None