dleerdefi/claude-code-construction / skills/rfi-drafter
rfi-drafter
VerifiedDraft RFIs and manage the ambient issue registry. Reviews issues surfaced by other skills, escalates to formal RFIs. Triggers: 'draft RFI', 'write RFI', 'drawing conflict', 'review issues', 'issue queue'.
| File | Kind | Size |
|---|---|---|
| SKILL.md | skill md | 12.2 KB |
| agents/openai.yaml | config | 260 B |
| references/issue-record.schema.json | config | 3.5 KB |
| references/issue-schema.md | markdown | 4.0 KB |
| references/quality-checks.md | markdown | 3.3 KB |
| references/research-checklist.md | markdown | 5.5 KB |
| references/rfi-format.md | markdown | 5.8 KB |
| scripts/rfi_export.py | script | 13.4 KB |
| scripts/shared.py | script | 554 B |
references/quality-checks.md
View on GitHub# Quality Checks — RFI Draft Validation
## Pre-Presentation Gates
Before presenting any RFI draft to the user, verify every item.
If any gate fails, fix before presenting — do not present a draft
with known deficiencies.
### Reference Specificity
- [ ] Every drawing reference includes sheet number AND grid location
or room number (never "the floor plan" or "the detail")
- [ ] Detail references use full format: "Detail 3/A-7.02"
- [ ] Spec references include section AND paragraph: "08 71 00, 2.01.A"
- [ ] Room references include number AND name: "Room 203 (Classroom)"
### Description Clarity
- [ ] A person unfamiliar with the project could understand the issue
from the description alone
- [ ] The description states what the documents show (facts), not
what the PE thinks is wrong (opinion)
- [ ] The conflict/ambiguity is explicitly identified — not implied
- [ ] No jargon without explanation (avoid abbreviations the A/E
might not use)
### Suggested Resolution
- [ ] A resolution is provided (never submit without one)
- [ ] Resolution is framed as suggestion, not directive
- [ ] Resolution includes reasoning (why this solution makes sense)
- [ ] Resolution is technically feasible (not just "fix it")
### Schedule Impact
- [ ] Specific construction activity affected is named
- [ ] Date of that activity is stated (or "TBD — verify with PM")
- [ ] Response-needed-by date is stated
- [ ] Impact is proportional (don't overstate for minor issues)
### Routing
- [ ] Addressee is the correct design professional for the scope
- [ ] If cross-discipline: addressed to Architect (prime), with
the relevant engineer CC'd
- [ ] From line matches the GC's project team
### Duplication
- [ ] No existing RFI addresses the same condition (checked log)
- [ ] If related RFI exists, it's referenced in the new draft
- [ ] Issue is not resolved by addenda, ASI, or general notes
### Document Precedence
- [ ] If the issue involves conflicting documents, the precedence
hierarchy is correctly applied and cited
- [ ] Large-scale vs small-scale drawing precedence noted if relevant
- [ ] Figured dimension vs scaled dimension precedence noted if
relevant
## Common Failure Patterns
These patterns indicate an RFI that will be returned or ignored:
**"Please clarify."** — Without stating what specifically needs
clarification. The A/E won't know what you're asking.
**"Per our discussion..."** — Verbal agreements must be formalized,
but the RFI must still state the technical issue independently.
**"The drawings are wrong."** — Presumptive. State what the drawings
show and why it creates a problem. Let the A/E determine if it's
an error.
**Missing drawing references.** — The A/E will need to look up the
exact location. Make their job easy — give them every reference.
**Scope creep.** — One issue per RFI. If you found three problems
in the same area, write three RFIs. Combined RFIs get partial
responses and create tracking nightmares.
## Post-Draft User Options
After presenting the quality-checked draft, offer:
1. **Export as .docx** — Produces a formatted Word document
2. **Refine** — User provides edits, Claude iterates
3. **Reject** — Issue isn't actually an issue; discard
4. **Defer** — Log back to issue registry for later action
5. **Batch** — User wants to draft multiple RFIs from the queue
Related skills
| Skill | Stars | Updated | Works with | Status | Install |
|---|---|---|---|---|---|
| reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis. | Stars: 471k | Updated | Claude CodeCodex | Verified | Install |
| diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness. | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependencies | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |
| executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is available | Stars: 322k | Updated | Claude CodeCodex | Verified | Install |