Decision Dashboard
All open decisions across the system. Sourced from FR “Open Questions” and “Decisions” sections. When decided: move to
logs/decision-log.mdand mark[x]in the FR.
Pending Decisions
| ID | Question | FR | Raised |
|---|---|---|---|
| PD-001 | Should session recap show open feature requests? | FR-004 | 2026-02-26 |
| PD-002 | Should a session-summary file be written at session end? | FR-004 | 2026-02-26 |
| PD-003 | Automate hallucination verification via hook? | FR-002 | 2026-02-26 |
| PD-004 | Add confidence level indicator to answers? | FR-002 | 2026-02-26 |
| PD-005 | Include “wins this week” in briefing? | FR-005 | 2026-02-27 |
| PD-006 | Persistent “current focus” marker — where and how? | FR-005 | 2026-02-27 |
| PD-007 | Which vault sync method? (Obsidian Sync / Git / Syncthing / Cloud) | FR-020 | 2026-02-26 |
| PD-008 | Which paths truly need protection? | FR-003 | 2026-02-26 |
| PD-009 | Tiered protection model (strict / soft / open) — worth the complexity? | FR-003 | 2026-02-26 |
| PD-010 | Wait with protection until post-MVP? | FR-003 | 2026-02-26 |
| PD-011 | Protected files hook: block or warn? | FR-003 | 2026-02-26 |
| PD-012 | Force override for protected files? | FR-003 | 2026-02-26 |
| PD-013 | Protected paths list: CLAUDE.md vs separate config? | FR-003 | 2026-02-26 |
| PD-014 | Workflow doc: one file or split into sections? | FR-030 | 2026-02-26 |
| PD-015 | Create a /workflow skill? | FR-030 | 2026-02-26 |
| PD-016 | Learning mode: auto-suggest next topic? | FR-062 | 2026-02-26 |
| PD-017 | Mini-lessons in daily briefings? | FR-062 | 2026-02-26 |
| PD-018 | How to measure if momentum support actually helps? | FR-005 | 2026-02-27 |
| PD-019 | FR-034 Phase 3: auto-categorization strategy for inbox items? | FR-034 | 2026-02-26 |
| PD-020 | Should Opus use the GitHub MCP server for richer API access? | FR-018 | 2026-02-26 |
| PD-021 | What gets pushed to GitHub vs what stays local (backup convention)? | FR-018 | 2026-02-26 |
| PD-022 | How to handle overlapping skills (e.g. /idea vs /new-feature)? | FR-017 | 2026-02-27 |
| PD-023 | Should existing .claude/rules/ files be removed after moving to path-scoped locations? | FR-026 | 2026-02-27 |
| PD-024 | Are there rules in CLAUDE.md that should be extracted into path-scoped rules? | FR-026 | 2026-02-27 |
| PD-025 | CLI framework choice for Python CLI entry point (Click vs argparse)? | FR-009 | 2026-02-27 |
| PD-026 | Where to document test conventions — vault/20_knowledge/tech/ or README? | FR-010 | 2026-02-27 |
| PD-027 | Use python-dotenv or alternative for loading env vars? | FR-015 | 2026-02-27 |
| PD-028 | Which hook has the highest value to implement first? | FR-011 | 2026-02-27 |
| PD-029 | Should verify-tests be agent-based or command-based (just runs test suite)? | FR-011 | 2026-02-27 |
| PD-030 | Should .claude/rules/** be protected, or just CLAUDE.md itself? | FR-003 | 2026-02-27 |
| PD-031 | Should the protect-files hook block edits or just warn? | FR-003 | 2026-02-27 |
| PD-032 | Start with bash scripts or Python for FR workflow automation Phase 1? | FR-028 | 2026-02-26 |
| PD-033 | Should feature request overview sync check auto-fix or just report mismatches? | FR-027 | 2026-02-27 |
| PD-034 | Should the sync check run as a hook after vault edits, or only on demand? | FR-027 | 2026-02-27 |
| PD-035 | Include a Mermaid flowchart in the workflow document? | FR-030 | 2026-02-26 |
| PD-036 | Error log format — single file or folder with daily files? | FR-049 | 2026-02-27 |
| PD-037 | Severity levels for error logging Phase 1 (simple vs structured)? | FR-049 | 2026-02-27 |
| PD-038 | When to switch from sequential (Phase 1) to scaffolding architecture (Phase 2) for briefings? | FR-037 | 2026-02-26 |
| PD-039 | Use Google Calendar MCP server for calendar integration in briefings? | FR-037 | 2026-03-04 |
| PD-040 | Claude Code manual structure — single file or multi-file? | FR-033 | 2026-02-26 |
| PD-041 | Use Fetch MCP server for scraping Claude Code docs? | FR-033 | 2026-02-26 |
| PD-042 | Which agent to build first (researcher, reviewer, or builder)? | FR-043 | 2026-02-27 |
| PD-043 | Agent definition format — own CLAUDE.md-style instructions or just prompt templates? | FR-043 | 2026-02-27 |
| PD-044 | How do agents interact with skills (e.g. briefing-compiler used by /briefing)? | FR-043 | 2026-02-27 |
| PD-045 | Per-agent memory — dedicated memory files per agent or shared memory pool? | FR-043 | 2026-03-04 |
| PD-046 | Persistent agent sessions — always-on server process or session-start preload from memory? | FR-043 | 2026-03-04 |
| PD-047 | Reviewer agent — hard gate (blocks build until approved) or advisory (warnings only)? | FR-043 | 2026-03-04 |
| PD-048 | Features specialist — auto-generate full spec from one-liner or require user outline? | FR-043 | 2026-03-04 |
| PD-049 | System self-review frequency (weekly / triggered / both)? | FR-048 | 2026-02-26 |
| PD-050 | Self-review scope — config only or include vault health? | FR-048 | 2026-02-26 |
| PD-051 | Health check trigger — /health skill, daily briefing, or both? | FR-048 | 2026-02-26 |
| PD-052 | Periodic system review frequency (weekly / after major changes / both)? | FR-048 | 2026-02-27 |
| PD-053 | Should Claude auto-trigger periodic reviews or only on user request? | FR-048 | 2026-02-27 |
| PD-054 | Periodic review scope — system docs only or include FR consistency? | FR-048 | 2026-02-27 |
| PD-055 | Vault maintenance frequency (daily / weekly)? | FR-032 | 2026-02-26 |
| PD-056 | validate-feature-request-format hook — PreToolUse or separate script? | FR-032 | 2026-02-26 |
| PD-057 | Which access method for Atlas vault — Filesystem MCP server vs direct path? | FR-025 | 2026-02-26 |
| PD-058 | Should the /learn-series skill generate 5-10 lessons at once? | FR-062 | 2026-02-26 |
| PD-059 | VPS provider selection (Hetzner / DigitalOcean / Contabo)? | FR-019 | 2026-02-26 |
| PD-060 | Chat bot platform for phone access (Telegram vs Discord vs other)? | FR-022 | 2026-02-26 |
| PD-061 | Email API choice — IMAP vs Gmail API? | FR-023 | 2026-02-26 |
| PD-062 | WhatsApp integration method — Business API vs whatsapp-web.js bridge? | FR-023 | 2026-02-26 |
| PD-063 | STT engine choice (Groq Whisper / OpenAI Whisper / local)? | FR-024 | 2026-02-26 |
| PD-064 | TTS engine choice (ElevenLabs / OpenAI TTS / alternatives)? | FR-024 | 2026-02-26 |
| PD-065 | Scope of autonomous implementation — which feature request types can be auto-built? | FR-066 | 2026-02-26 |
| PD-066 | Nano Banana model/tier to use (free vs Pro)? | FR-071 | 2026-03-04 |
| PD-067 | Access Nano Banana via Gemini API directly or through a wrapper? | FR-071 | 2026-03-04 |
| PD-068 | Primary feature requests for image generation (vault illustrations / social media / personal)? | FR-071 | 2026-03-04 |
| PD-069 | External repo research scope — Claude Code repos only, or also similar AI-assistant setups? | FR-067 | 2026-03-04 |
| PD-070 | External repo research trigger — cron schedule or on-demand only? | FR-067 | 2026-03-04 |
| PD-071 | Kanban board format — static markdown or interactive HTML? | FR-038 | 2026-03-04 |
| PD-072 | Kanban board generation — Claude skill or auto-generated via script? | FR-038 | 2026-03-04 |
| PD-073 | Personal task management: what Trello functionality to replicate first? | FR-068 | 2026-03-04 |
| PD-074 | Personal task management: Obsidian-native or separate tool? | FR-068 | 2026-03-04 |
| PD-075 | Personal task management: how to handle recurring tasks? | FR-068 | 2026-03-04 |
| PD-076 | Personal task management: should tasks connect to rest of Opus from day one? | FR-068 | 2026-03-04 |
| PD-077 | Proposal categories — which category list to use (8 types / 4 simplified / tags only)? | FR-036 | 2026-03-04 |
| PD-078 | Proposal storage — subfolders per category or tags in frontmatter? | FR-036 | 2026-03-04 |
| PD-079 | Template review: which new templates are needed (idea+dashboard / all three / none)? | FR-001 | 2026-03-04 |
| PD-080 | Should templates have a version field? | FR-001 | 2026-03-04 |
| PD-081 | Task scheduler runtime approach — Python daemon / SQLite polling loop / system cron? | FR-044 | 2026-03-04 |
| PD-082 | Task registration format — SQLite / config file (YAML) / vault markdown? | FR-044 | 2026-03-04 |
| PD-083 | Where does the scheduler run first — local or VPS-only? | FR-044 | 2026-03-04 |
| PD-084 | Complexity scoring factors and weights for go-command routing? | FR-045 | 2026-03-04 |
| PD-085 | Threshold boundaries between single-agent, pipeline, and team strategies (go-command)? | FR-045 | 2026-03-04 |
| PD-086 | What happens when a dispatched agent gets stuck — escalate / retry / hand off? | FR-045 | 2026-03-04 |
| PD-087 | Should /go require explicit user confirmation before dispatching? | FR-045 | 2026-03-04 |
| PD-088 | Job registry storage backend — JSON / SQLite / in-memory? | FR-046 | 2026-03-04 |
| PD-089 | Maximum concurrent agents in the job queue? | FR-046 | 2026-03-04 |
| PD-090 | Job priority formula — fixed categories with aging or weighted scoring? | FR-046 | 2026-03-04 |
| PD-091 | Starvation prevention threshold — how quickly do low-priority jobs escalate? | FR-046 | 2026-03-04 |
| PD-092 | Workflow state machine notification channel — terminal only / vault inbox / messaging? | FR-031 | 2026-03-04 |
| PD-093 | Stale threshold — how many days without activity before flagging a feature request? | FR-031 | 2026-03-04 |
| PD-094 | Gate enforcement — strict (blocks transition) or advisory (warns but allows)? | FR-031 | 2026-03-04 |
| PD-095 | Auto-close — fully automatic on user confirmation or require explicit archive command? | FR-031 | 2026-03-04 |
| PD-096 | Pitch state — include in state machine or handle separately in FR-050? | FR-031 | 2026-03-04 |
| PD-097 | Proactive monitoring frequency — hourly / daily / event-driven / mixed? | FR-050 | 2026-03-04 |
| PD-098 | Which external sources to scan first — GitHub releases / RSS / Reddit? | FR-050 | 2026-03-04 |
| PD-099 | Pitch delivery channel — terminal only / vault inbox / messaging? | FR-050 | 2026-03-04 |
| PD-100 | On pitch approval: auto-create draft FR or add to inbox for manual processing? | FR-050 | 2026-03-04 |
| PD-101 | Self-improvement scope — can the system modify its own code or only propose changes? | FR-050 | 2026-03-04 |
| PD-102 | Default stale thresholds per lifecycle status (conservative / aggressive / user-configurable)? | FR-040 | 2026-03-04 |
| PD-103 | How many nudges before escalating a stale FR? | FR-040 | 2026-03-04 |
| PD-104 | Auto-advance/cancel rules — always require user confirmation or configurable per rule? | FR-040 | 2026-03-04 |
| PD-105 | Stale nudge delivery channel — dashboard only / briefing / push via Telegram? | FR-040 | 2026-03-04 |
| PD-106 | Priority scoring factor weights — stage-heavy / balanced / blocker-heavy? | FR-039 | 2026-03-04 |
| PD-107 | Priority score stored in frontmatter, computed on-the-fly, or both? | FR-039 | 2026-03-04 |
| PD-108 | Manual priority override handling — additive modifier / full override / pin? | FR-039 | 2026-03-04 |
| PD-109 | Should feature request dashboard sort by priority score by default? | FR-039 | 2026-03-04 |
| PD-110 | Should /approve allow skipping lifecycle stages? | FR-029 | 2026-03-04 |
| PD-111 | Should rejected FRs go to rejected/ folder or back to ideas/? | FR-029 | 2026-03-04 |
| PD-112 | Should /implement auto-start work or generate a plan for review first? | FR-029 | 2026-03-04 |
| PD-113 | How does /implement interact with complexity routing (FR-045)? | FR-029 | 2026-03-04 |
| PD-114 | Should complexity scoring be heuristic-only, LLM-assisted, or hybrid? | FR-045 | 2026-03-04 |
| PD-115 | What are the exact complexity tier threshold boundaries (simple / medium / complex)? | FR-045 | 2026-03-04 |
| PD-116 | Should the quality gate confidence threshold be configurable? | FR-045 | 2026-03-04 |
| PD-117 | How to handle disagreement between advocate and reviewer agents? | FR-045 | 2026-03-04 |
| PD-118 | Obsidian sync method — Obsidian Sync / Syncthing / Git (related to FR-020)? | FR-021 | 2026-03-04 |
| PD-119 | Obsidian mobile app — Obsidian Mobile (iOS/Android) or view via synced files only? | FR-021 | 2026-03-04 |
| PD-120 | Obsidian plugin scope — minimal or full power-user setup? | FR-021 | 2026-03-04 |
| PD-121 | Workflow Control blocks — Buttons plugin or custom Obsidian plugin? | FR-021 | 2026-03-04 |
121 open decisions.