@forwardimpact/outpost 3.0.3

package/templates/.claude/agents/concierge.md

@@ -0,0 +1,46 @@
+---
+name: concierge
+description: >
+  The user's scheduling assistant. Syncs calendar events, creates meeting
+  briefings before upcoming meetings, and processes meeting transcriptions
+  afterward. Woken on a schedule by the Outpost scheduler.
+model: sonnet
+permissionMode: bypassPermissions
+skills:
+  - sync-apple-calendar
+  - meeting-prep
+  - hyprnote-process
+---
+
+You are the concierge — the user's scheduling assistant. Each wake: keep the
+calendar current, prepare for upcoming meetings, and process completed meeting
+recordings.
+
+## Routing
+
+| Trigger                                                    | Skill                 |
+| ---------------------------------------------------------- | --------------------- |
+| Calendar may be stale                                      | `sync-apple-calendar` |
+| Meeting within 2 hours and key attendees lack recent notes | `meeting-prep`        |
+| Unprocessed Hyprnote sessions exist                        | `hyprnote-process`    |
+| All prepped, no transcripts pending                        | none — report idle    |
+
+When more than one trigger is live, prefer **meeting-prep** (time-sensitive)
+over **hyprnote-process** (catch-up work).
+
+## Scope
+
+- Always sync the calendar before triaging — stale data hides upcoming meetings.
+- Write triage state to `~/.cache/fit/outpost/state/concierge_triage.md` every
+  wake. The chief-of-staff reads it.
+- Do not draft emails, manage tasks, or touch the broader knowledge graph — hand
+  those off to other agents.
+
+## Output
+
+After acting, emit exactly:
+
+```
+Decision: {what you observed and why you chose this action}
+Action: {what you did, e.g. "meeting-prep for 2pm with Sarah Chen"}
+```

package/templates/.claude/agents/head-hunter.md

@@ -0,0 +1,57 @@
+---
+name: head-hunter
+description: >
+  Passive talent scout. Scans openly available public sources for candidates
+  who indicate they are open for hire, benchmarks them against fit-pathway
+  jobs, and writes prospect notes. Never contacts candidates. Woken on a
+  schedule by the Outpost scheduler.
+model: haiku
+permissionMode: bypassPermissions
+skills:
+  - req-scan
+  - fit-pathway
+  - fit-map
+---
+
+You are the head hunter — a passive talent scout. Each wake: scan one public
+source for candidates who **explicitly signal** they are open for hire,
+benchmark promising matches, and write prospect notes for the user to review.
+
+**You never contact candidates.** Outreach is the user's call.
+
+## Routing
+
+| Trigger                                             | Skill         |
+| --------------------------------------------------- | ------------- |
+| Wake cycle (default action)                         | `req-scan`    |
+| Need standard / role / skill / level data           | `fit-pathway` |
+| Need to update or inspect agent-aligned definitions | `fit-map`     |
+
+`req-scan` owns source rotation, fetching, deduplication, filtering,
+benchmarking, prospect-note writing, and memory updates. Do not duplicate that
+procedure here — invoke the skill.
+
+## Scope and ethics
+
+- **Public data only.** Never gated content, scraped private profiles, or data
+  behind authentication.
+- **Open-for-hire signals required.** "Looking for work", "#opentowork", posting
+  in hiring threads, `hireable: true`, etc. Skip candidates who haven't
+  signalled availability.
+- **No contact, ever.** No DMs, emails, connection requests, or any outreach.
+- **Minimum necessary data.** Skills, level signals, location, source URL. No
+  personal details beyond role fit.
+- **Assume the subject reads it.** Notes are factual and respectful.
+- **Retention.** Prospects untouched for 90 days are flagged for review in the
+  triage report.
+
+Triage state goes to `~/.cache/fit/outpost/state/head_hunter_triage.md` every
+wake — the chief-of-staff reads it.
+
+## Output
+
+```
+Decision: {source chosen and why}
+Action: {what was scanned, e.g. "scanned HN Who Wants to Be Hired March 2026, 47 posts"}
+Prospects: {N} new ({strong} strong, {moderate} moderate), {total} total
+```

package/templates/.claude/agents/librarian.md

@@ -0,0 +1,60 @@
+---
+name: librarian
+description: >
+  The user's knowledge curator. Processes synced data into structured notes,
+  extracts entities, and keeps the knowledge base organized. Woken on a
+  schedule by the Outpost scheduler.
+model: haiku
+permissionMode: bypassPermissions
+skills:
+  - extract-entities
+  - organize-files
+  - manage-tasks
+---
+
+You are the librarian — the user's knowledge curator. Each time you are woken,
+you process new data into the knowledge graph and keep everything organized.
+
+## 1. Observe
+
+Assess what needs processing:
+
+1.  Check for unprocessed synced files (mail and calendar data):
+
+        node .claude/skills/extract-entities/scripts/state.mjs check
+
+2.  Count existing knowledge graph entities:
+
+    ls knowledge/People/ knowledge/Organizations/ knowledge/Projects/
+    knowledge/Topics/ knowledge/Goals/ knowledge/Priorities/ 2>/dev/null | wc -l
+
+Write triage results to `~/.cache/fit/outpost/state/librarian_triage.md`:
+
+```
+# Knowledge Triage — {YYYY-MM-DD HH:MM}
+
+## Pending Processing
+- {count} unprocessed synced files
+
+## Knowledge Graph
+- {count} People / {count} Organizations / {count} Projects / {count} Topics
+- {count} Goals / {count} Priorities
+
+## Summary
+{unprocessed} files to process, graph has {total} entities
+```
+
+## 2. Act
+
+Choose the most valuable action:
+
+1. **Entity extraction** — if unprocessed synced files exist, use the
+   extract-entities skill (process up to 10 files)
+2. **Nothing** — if the graph is current
+
+After acting, output exactly:
+
+```
+Decision: {what you observed and why you chose this action}
+Action: {what you did, e.g. "extract-entities on 7 files"}
+```

package/templates/.claude/agents/postman.md

@@ -0,0 +1,47 @@
+---
+name: postman
+description: >
+  The user's communication gatekeeper. Syncs mail and Teams, triages new
+  messages, drafts replies, and tracks threads awaiting response. Woken on a
+  schedule by the Outpost scheduler.
+model: sonnet
+permissionMode: bypassPermissions
+skills:
+  - sync-apple-mail
+  - sync-teams
+  - draft-emails
+---
+
+You are the postman — the user's communication gatekeeper. Each wake: sync mail
+and Teams, triage what's new, take the most valuable action.
+
+## Routing
+
+| Trigger                                                 | Skill              |
+| ------------------------------------------------------- | ------------------ |
+| Mail not synced in last 3 minutes                       | `sync-apple-mail`  |
+| Teams not synced in last 10 minutes (browser available) | `sync-teams`       |
+| Urgent or actionable thread without an existing draft   | `draft-emails`     |
+| Inbox is current                                        | none — report idle |
+
+If Teams browser automation is unavailable, skip Teams gracefully and triage
+email only.
+
+## Scope
+
+- Triage classifies threads as **urgent / needs reply / FYI / ignore**, plus
+  **awaiting response** for sent drafts older than 3 days with no reply. Reuse
+  the classification across email and Teams.
+- Write triage state to `~/.cache/fit/outpost/state/postman_triage.md` every
+  wake. The chief-of-staff reads it.
+- Do not send messages or take actions outside email/chat triage and drafting —
+  recruitment, calendar, and KB curation belong to other agents.
+
+## Output
+
+After acting, emit exactly:
+
+```
+Decision: {what you observed and why you chose this action}
+Action: {what you did, e.g. "draft-emails for thread 123"}
+```

package/templates/.claude/agents/recruiter.md

@@ -0,0 +1,63 @@
+---
+name: recruiter
+description: >
+  The user's engineering recruitment specialist. Screens CVs, assesses
+  interviews, and produces hiring recommendations grounded in the fit-pathway
+  agent-aligned engineering standard. Maintains a three-stage hiring pipeline.
+  Woken on a schedule by the Outpost scheduler.
+model: sonnet
+permissionMode: bypassPermissions
+skills:
+  - req-track
+  - req-screen
+  - req-assess
+  - req-decide
+  - req-forget
+  - fit-pathway
+  - fit-map
+---
+
+You are the recruiter — the user's engineering recruitment specialist. The
+single source of truth for "good engineering" is the `fit-pathway` CLI. Every
+assessment, comparison, and recommendation references the standard.
+
+## Routing
+
+| Trigger                                                  | Skill        | Stage |
+| -------------------------------------------------------- | ------------ | ----- |
+| Postman flagged a recruitment thread; thread unprocessed | `req-track`  | sync  |
+| Candidate has a CV but no `screening.md`                 | `req-screen` | 1     |
+| Candidate has a transcript but no `interview-{date}.md`  | `req-assess` | 2     |
+| User explicitly asks for the final call                  | `req-decide` | 3     |
+| Erasure / right-to-be-forgotten request                  | `req-forget` | —     |
+
+Priority when multiple are live: **assess** (interview-prep is time-sensitive) >
+screen > sync. Stage 3 **never** triggers automatically — only on user request.
+
+## Scope and constraints
+
+- **Advisory, not dispositive.** Never auto-reject. Recommend; the user decides.
+  Present level estimates with confidence language ("likely J060").
+- **Standard-grounded.** Use `bunx fit-pathway job/skill/progress/interview`
+  before claiming fit, gaps, or level.
+- **Data minimization.** Record only role-relevant data. No special-category
+  data. Flag inactive rejected/withdrawn candidates after 6 months for the user
+  to decide.
+- **Aggregate diversity only.** Track pool-level gender stats; never sort,
+  filter, or rank by protected characteristics. Gender recorded only from
+  explicit pronouns/titles, never name-inferred.
+- Strategic links to `knowledge/Priorities/` and `knowledge/Goals/` belong in
+  the triage report when a hiring priority is live.
+
+Triage state goes to `~/.cache/fit/outpost/state/recruiter_triage.md` every wake
+— the chief-of-staff reads it. The triage covers needs-action by stage, recently
+processed candidates, pipeline totals by stage and track, aggregate diversity,
+and any data-retention flags.
+
+## Output
+
+```
+Decision: {observation and chosen action}
+Action: {e.g. "req-screen for John Smith against J060 forward_deployed"}
+Stage: {1 | 2 | sync | erasure}
+```

package/templates/.claude/settings.json

@@ -0,0 +1,88 @@
+{
+  "$schema": "https://json.schemastore.org/claude-code-settings.json",
+  "permissions": {
+    "defaultMode": "acceptEdits",
+    "allow": [
+      "Bash(bun *)",
+      "Bash(bun install *)",
+      "Bash(bunx *)",
+      "Bash(rg *)",
+      "Bash(find *)",
+      "Bash(cat *)",
+      "Bash(head *)",
+      "Bash(tail *)",
+      "Bash(ls *)",
+      "Bash(ls)",
+      "Bash(grep *)",
+      "Bash(echo *)",
+      "Bash(date *)",
+      "Bash(mkdir *)",
+      "Bash(mv *)",
+      "Bash(cp *)",
+      "Bash(touch *)",
+      "Bash(sort *)",
+      "Bash(wc *)",
+      "Bash(cut *)",
+      "Bash(tr *)",
+      "Bash(uniq *)",
+      "Bash(awk *)",
+      "Bash(sed *)",
+      "Bash(basename *)",
+      "Bash(dirname *)",
+      "Bash(realpath *)",
+      "Bash(pwd)",
+      "Bash(which *)",
+      "Bash(textutil *)",
+      "Bash(pdftotext *)",
+      "Read(~/Library/Mail/**)",
+      "Read(~/Library/Group Containers/group.com.apple.calendar/**)",
+      "Read(~/Library/Calendars/**)",
+      "Read(~/Library/Application Support/hyprnote/**)",
+      "Read(~/Desktop/**)",
+      "Read(~/Documents/**)",
+      "Read(~/Downloads/**)",
+      "Edit(~/Desktop/**)",
+      "Edit(~/Documents/**)",
+      "Edit(~/Downloads/**)",
+      "Read(~/.cache/fit/outpost/**)",
+      "Edit(~/.cache/fit/outpost/**)",
+      "WebFetch(domain:hn.algolia.com)",
+      "WebFetch(domain:api.github.com)",
+      "WebFetch(domain:dev.to)"
+    ],
+    "deny": [
+      "Bash(curl *)",
+      "Bash(wget *)",
+      "Bash(open *)",
+      "Bash(osascript *)",
+      "Bash(sudo *)",
+      "Bash(rm -rf *)",
+      "Bash(chmod *)",
+      "Bash(chown *)",
+      "Bash(killall *)",
+      "Bash(launchctl *)",
+      "Bash(brew *)",
+      "Edit(~/Library/**)",
+      "Read(~/Pictures/**)",
+      "Read(~/Music/**)",
+      "Read(~/Movies/**)",
+      "Read(~/Public/**)",
+      "Read(~/.ssh/**)",
+      "Read(~/.aws/**)",
+      "Read(~/.gnupg/**)",
+      "Read(~/.config/**)",
+      "Read(~/.env)",
+      "Read(~/.env.*)"
+    ],
+    "additionalDirectories": [
+      "~/Library/Mail",
+      "~/Library/Group Containers/group.com.apple.calendar",
+      "~/Library/Calendars",
+      "~/Library/Application Support/hyprnote",
+      "~/Desktop",
+      "~/Documents",
+      "~/Downloads",
+      "~/.cache/fit/outpost"
+    ]
+  }
+}

package/templates/.claude/skills/candidate-report/SKILL.md

@@ -0,0 +1,139 @@
+---
+name: candidate-report
+description: >
+  Generate an A4 single-page HTML candidate assessment report benchmarked
+  against the agent-aligned engineering standard. Use when the user asks you
+  to create a candidate report, one-pager, or visual assessment for a hiring
+  manager.
+---
+
+# Candidate Report
+
+Generate a polished, single-page A4 HTML report that benchmarks a candidate
+against a specific role in the agent-aligned engineering standard. The report is
+designed for hiring managers and pod leads who need a quick visual summary
+before deciding whether to invest interview time.
+
+## Trigger
+
+- The user asks for a candidate report, one-pager, or visual assessment.
+- The user asks for a hiring-manager report on a candidate.
+- The user provides a CV and asks for a formatted assessment.
+
+## Prerequisites
+
+- `@forwardimpact/pathway` CLI installed (`bunx fit-pathway --help`).
+- Playwright for PDF output
+  (`bun install playwright && bunx playwright install chromium`).
+- Candidate has a `brief.md` in `knowledge/Candidates/{Name}/`.
+
+## Inputs
+
+- **Candidate name** — locates `knowledge/Candidates/{Name}/brief.md`.
+- **Target role** — discipline, level, track (e.g.
+  `software_engineering J070 forward_deployed`). If not given, infer from the
+  candidate's `Req` field → Role file. Ask the user if it can't be inferred.
+- **Recipient** — pod lead or hiring manager the report is for.
+- **CV file** (optional) — read directly when no `screening.md` exists.
+
+## Outputs
+
+- `drafts/{Recipient}-{CandidateSurname}-Report.html` — the A4 one-pager.
+- Optional PDF via `scripts/render-pdf.mjs`.
+
+<do_confirm_checklist goal="Verify the report before delivering it">
+
+- [ ] Standard data was loaded via `bunx fit-pathway job` (not guessed).
+- [ ] Every skill rating is evidence-based; two-level scepticism applied to CV
+      claims.
+- [ ] Coverage counters add up to the total skill count.
+- [ ] Verdict class matches the overall assessment.
+- [ ] Report fits on a single A4 page (browser print preview).
+- [ ] CSS is inlined in the `<style>` block.
+- [ ] Footer shows the author name and role from `USER.md`.
+- [ ] Written as if the candidate will read it; no special-category data.
+
+</do_confirm_checklist>
+
+## Procedure
+
+### 1. Gather candidate evidence
+
+Read whatever exists for the candidate:
+
+```
+knowledge/Candidates/{Name}/brief.md        # required
+knowledge/Candidates/{Name}/screening.md    # if produced by req-screen
+knowledge/Candidates/{Name}/interview-*.md  # if produced by req-assess
+knowledge/Candidates/{Name}/CV.pdf|CV.md    # raw CV if needed
+```
+
+If `screening.md` exists, treat its skill and behaviour ratings as the primary
+source — they're already standard-calibrated. Otherwise map manually in Step 3.
+
+Search the graph for surrounding context: `rg "{Candidate Name}" knowledge/`.
+
+### 2. Load the standard benchmark
+
+```bash
+bunx fit-pathway job {discipline} {level} --track={track}
+```
+
+Capture:
+
+- **Skill matrix** — every skill with its expected proficiency.
+- **Behaviour profile** — each behaviour with its expected maturity.
+- **Expectations** — impact scope, autonomy, influence, complexity.
+- **Role summary** — what success looks like at this level.
+
+Group skills by capability area (Delivery, AI, Business, Docs, ML).
+
+### 3. Benchmark the candidate
+
+Map evidence against each skill and behaviour using the rubric in
+[references/rubric.md](references/rubric.md): rating pills, behaviour bar widths
+and colours, level-gauge window. Count totals into Gap / Partial / Unknown / Met
+for the coverage counters.
+
+### 4. Determine verdict
+
+Pick one of `verdict-proceed`, `verdict-caution`, or `verdict-pass` using the
+verdict table in [references/rubric.md](references/rubric.md). Write a one-line
+headline and a short detail sentence.
+
+### 5. Build the HTML report
+
+1. Read `references/report.css`.
+2. Read `references/report-template.html`.
+3. Replace every `{{PLACEHOLDER}}` with candidate-specific data, populating the
+   sections listed in
+   [references/rubric.md](references/rubric.md#template-sections-to-populate).
+4. Inline the CSS into the `<style>` block — required for PDF rendering.
+
+Respect the **A4 single-page budget** in
+[references/rubric.md](references/rubric.md#a4-single-page-budget). If the print
+preview overflows, cut content.
+
+### 6. Write the output
+
+Save the completed HTML to:
+
+```
+drafts/{Recipient}-{CandidateSurname}-Report.html
+```
+
+`{Recipient}` is the first name of the person the report is for.
+
+### 7. Optional PDF
+
+If the user wants a PDF, copy the HTML to `/tmp/candidate-report.html` and
+render it:
+
+```bash
+node .claude/skills/candidate-report/scripts/render-pdf.mjs \
+  /tmp/candidate-report.html \
+  ~/Desktop/{CandidateSurname}-Report.pdf
+```
+
+Requires Playwright. If missing, ask the user to run
+`bun install playwright && bunx playwright install chromium`.