chore: rename job-hunt to pupila across code, scripts, and env vars

.claude/skills/pupila-ai-review/SKILL.md

@@ -2,10 +2,10 @@
 name: pupila-ai-review
 description: How the AI per-job review pipeline works in this repo - generating verdicts via local LLM CLI, parsing markdown-fenced JSON, the candidate brief lever, and the AI Apply tailored-package flow. Use when modifying src/ai-review.ts, src/lib/ai-apply.ts, tuning verdict prompts, debugging review output, editing the candidate brief, or wiring a new LLM provider.
 metadata:
-  scope: pupila / job-hunt
+  scope: pupila
 ---
 
-`pnpm run ai-review` is a **local-only** companion that augments selected jobs with an LLM review via `src/lib/llm.ts` (auto-detects `claude` / `codex` / `gemini` / `opencode`, override `JOB_HUNT_LLM`). Uses the user's local subscription (e.g. Claude Max) — NOT an API key, so no per-token charges.
+`pnpm run ai-review` is a **local-only** companion that augments selected jobs with an LLM review via `src/lib/llm.ts` (auto-detects `claude` / `codex` / `gemini` / `opencode`, override `PUPILA_LLM`). Uses the user's local subscription (e.g. Claude Max) — NOT an API key, so no per-token charges.
 
 The launchd/cron review agent runs daily at 07:15 by default. Without an LLM CLI, run `scripts/install-launchd.sh --no-review` (or cron equivalent).
 
@@ -54,7 +54,7 @@ Tests in `tests/ai-review-parse.test.ts` (9 cases) cover all the failure modes s
 
 ## The candidate brief
 
-`config/candidate-brief.md` is the **only natural-language config in the repo**. Hand-edited, gitignored. Generated via `pnpm run setup-brief --file ~/cv.pdf` (or via the UI's Profile tab → drop a PDF/DOCX/MD CV). The CLI shells out to whichever local LLM CLI is installed (`claude` / `codex` / `gemini` / `opencode` — auto-detected; override via `JOB_HUNT_LLM=<provider>`).
+`config/candidate-brief.md` is the **only natural-language config in the repo**. Hand-edited, gitignored. Generated via `pnpm run setup-brief --file ~/cv.pdf` (or via the UI's Profile tab → drop a PDF/DOCX/MD CV). The CLI shells out to whichever local LLM CLI is installed (`claude` / `codex` / `gemini` / `opencode` — auto-detected; override via `PUPILA_LLM=<provider>`).
 
 The brief is embedded **verbatim** in the review prompt — it's the main lever for tuning match/skip behaviour. To change verdicts at scale, edit the brief; to change individual scores, see the `pupila-filters` skill.
 

.claude/skills/pupila-fetchers/SKILL.md

@@ -2,7 +2,7 @@
 name: pupila-fetchers
 description: How to add a new job-source fetcher (ATS API, RSS, scraper) or extend tier-S slug lists in this repo. Use when adding a new job board, integrating a new ATS, scraping a new careers site, registering a new company under Ashby/Greenhouse/Lever, or diagnosing a fetcher that returned zero items.
 metadata:
-  scope: pupila / job-hunt
+  scope: pupila
 ---
 
 The pipeline ingests from 13 public sources (3 ATS APIs + RSS, JSON boards, HN, HTML scrapers, an Aave Next.js scraper, and `ashby-private` for orgs whose public posting-API is disabled). Adding a source means: a fetcher, a normalizer, a `Source` literal, a slot in the orchestrator, and dedup/render wiring.

.claude/skills/pupila-filters/SKILL.md

@@ -2,7 +2,7 @@
 name: pupila-filters
 description: How to tune job filter scoring, hard-drop rules, or debug why a specific job was kept/dropped via _signals. Use when adjusting weights in config/profile.json, adding a hard-exclude rule, tuning keyword lists, debugging fitScore, or interpreting the per-job _signals breakdown.
 metadata:
-  scope: pupila / job-hunt
+  scope: pupila
 ---
 
 All filter logic lives in `src/filters.ts`. Weights + keyword lists load from `config/profile.json` at runtime via `loadProfile()` (NOT a static import — the file is gitignored and auto-bootstrapped from `config/profile.default.json` on first run). Adjusting weights or keywords is a **non-code edit** to `profile.json`.

AGENTS.md

@@ -8,7 +8,7 @@ Guidance for future Codex sessions working in this repo.
 
 The repo ships a **neutral template** in `config/profile.json`. After onboarding (CV upload → brief generation), `/api/profile-generate` shells out to the local LLM CLI to fill in the personal keyword lists + weights based on the brief. Re-runnable from Settings → Scoring profile → Regenerate. `config/slugs.json` ships with the full ~50-company tier-S list (all public ATS URLs — non-personal data, edit by hand to add/remove companies).
 
-**First-run UX**: a forker generates their `config/candidate-brief.md` by running `pnpm run setup-brief --file ~/cv.pdf` (or via the UI's Profile tab → drop a PDF/DOCX/MD CV). That CLI shells out to whichever local LLM CLI is installed (`Codex`, `codex`, `gemini`, `opencode` — auto-detected, override via `JOB_HUNT_LLM=<provider>`).
+**First-run UX**: a forker generates their `config/candidate-brief.md` by running `pnpm run setup-brief --file ~/cv.pdf` (or via the UI's Profile tab → drop a PDF/DOCX/MD CV). That CLI shells out to whichever local LLM CLI is installed (`Codex`, `codex`, `gemini`, `opencode` — auto-detected, override via `PUPILA_LLM=<provider>`).
 
 ## Stack
 
@@ -279,7 +279,7 @@ The read happens **after** filter+dedup+sort but **before** `writeJson('data/job
 
 ## RSS feed
 
-[`src/feed.ts`](./src/feed.ts) emits a hand-rolled RSS 2.0 XML to `data/feed.xml` containing the top 50 `newJobs` by `fitScore`. The XML is hand-built (not via fast-xml-parser) because we control the content shape — `escapeXml` covers the five entity classes. The feed metadata (title, description, link) is overridable via `JOB_HUNT_FEED_TITLE` / `JOB_HUNT_FEED_DESC` / `JOB_HUNT_FEED_LINK` env vars. To subscribe locally, point your RSS reader at the `file://` path of `data/feed.xml`. (No remote URL anymore — the project is local-first.)
+[`src/feed.ts`](./src/feed.ts) emits a hand-rolled RSS 2.0 XML to `data/feed.xml` containing the top 50 `newJobs` by `fitScore`. The XML is hand-built (not via fast-xml-parser) because we control the content shape — `escapeXml` covers the five entity classes. The feed metadata (title, description, link) is overridable via `PUPILA_FEED_TITLE` / `PUPILA_FEED_DESC` / `PUPILA_FEED_LINK` env vars. To subscribe locally, point your RSS reader at the `file://` path of `data/feed.xml`. (No remote URL anymore — the project is local-first.)
 
 ## Salary parsing
 
@@ -345,7 +345,7 @@ The HTML has `<meta name="robots" content="noindex,nofollow">` as belt-and-suspe
 
 ## AI per-job review (`pnpm run ai-review`)
 
-[`src/ai-review.ts`](./src/ai-review.ts) is a **local-only** companion to the daily pipeline that augments selected jobs with an LLM review. It shells out via `src/lib/llm.ts` (auto-detects `Codex` / `codex` / `gemini` / `opencode` on PATH, override with `JOB_HUNT_LLM=<provider>`). The CLI uses the user's local subscription (e.g. Codex Max) — **not** an API key, so there are no per-token charges. With the new local-first scheduling, the launchd/cron review agent runs this every day at 07:15 by default. If you don't have an LLM CLI installed, run `scripts/install-launchd.sh --no-review` (or the cron equivalent) to skip the review step.
+[`src/ai-review.ts`](./src/ai-review.ts) is a **local-only** companion to the daily pipeline that augments selected jobs with an LLM review. It shells out via `src/lib/llm.ts` (auto-detects `Codex` / `codex` / `gemini` / `opencode` on PATH, override with `PUPILA_LLM=<provider>`). The CLI uses the user's local subscription (e.g. Codex Max) — **not** an API key, so there are no per-token charges. With the new local-first scheduling, the launchd/cron review agent runs this every day at 07:15 by default. If you don't have an LLM CLI installed, run `scripts/install-launchd.sh --no-review` (or the cron equivalent) to skip the review step.
 
 **Inputs:**
 - `data/jobs.json` — the slim list (committed)

CLAUDE.md

@@ -20,8 +20,8 @@ Guidance for future Claude Code sessions working in this repo. **Slim by design*
 Cross-cutting invariants (apply repo-wide):
 
 - **`config/profile.json` is gitignored** — encodes personal scoring preferences. Auto-bootstraps from committed `config/profile.default.json` on first `pnpm run dev` / `pnpm run ui` via `bootstrapProfileIfMissing()` (idempotent — `COPYFILE_EXCL` no-ops on the steady state). Don't bypass; don't commit personalized weights.
-- **Mandatory CV gate**: `pnpm run dev` checks for `config/candidate-brief.md` at startup and exits 1 if missing. Bypass with `JOB_HUNT_NO_BRIEF_CHECK=1` or `--no-brief-check`.
-- **`config/candidate-brief.md` is the only natural-language config** (gitignored). Generated via `pnpm run setup-brief --file ~/cv.pdf` or via the UI's Profile tab (drop PDF/DOCX/MD CV). CLI shells out to `claude`/`codex`/`gemini`/`opencode` — auto-detected, override `JOB_HUNT_LLM=<provider>`.
+- **Mandatory CV gate**: `pnpm run dev` checks for `config/candidate-brief.md` at startup and exits 1 if missing. Bypass with `PUPILA_NO_BRIEF_CHECK=1` or `--no-brief-check`.
+- **`config/candidate-brief.md` is the only natural-language config** (gitignored). Generated via `pnpm run setup-brief --file ~/cv.pdf` or via the UI's Profile tab (drop PDF/DOCX/MD CV). CLI shells out to `claude`/`codex`/`gemini`/`opencode` — auto-detected, override `PUPILA_LLM=<provider>`.
 - **Local-first scheduling**: daily aggregation runs via `scripts/install-launchd.sh` (macOS) or `scripts/install-cron.sh` (Linux), not GitHub Actions cron. CI runs only on push/PR for gates.
 - **`data/applied.json` source of truth** for application tracking (UI writes via Vite middleware). Commit manually to persist across machines. **Don't filter applied jobs out of the main list** — user explicitly wants them visible.
 
@@ -79,7 +79,7 @@ pnpm run mcp                          # MCP server over stdio
 
 `pnpm run dev` → `tsx src/index.ts` is the main pipeline (what launchd/cron runs). Steps:
 
-1. **CV gate** — fail-fast if `config/candidate-brief.md` missing (bypass: `JOB_HUNT_NO_BRIEF_CHECK=1` or `--no-brief-check`).
+1. **CV gate** — fail-fast if `config/candidate-brief.md` missing (bypass: `PUPILA_NO_BRIEF_CHECK=1` or `--no-brief-check`).
 2. **Profile bootstrap** — `bootstrapProfileIfMissing()` copies `config/profile.default.json` → `profile.json` on first run.
 3. **Fetch** — all 13 sources in parallel via `processFetcher()` + `Promise.all`. Each fetcher returns `{ items, errors }` and **never throws** (a rejection would kill the whole run).
 4. **Normalize** — per-source `normalize<Source>()` → `Job[]`. Salary fields populated via `withSalary()` spread.
@@ -134,7 +134,7 @@ Settings tab (eight panels), Jinder (swipe-to-apply queue), AI Apply (per-job ta
 
 ## AI per-job review
 
-`pnpm run ai-review` is a **local-only** companion that augments selected jobs with an LLM review via `src/lib/llm.ts` (auto-detects `claude`/`codex`/`gemini`/`opencode`, override `JOB_HUNT_LLM`). Uses the local subscription — **not** an API key, so no per-token charges. Output: `data/ai-reviews.json`.
+`pnpm run ai-review` is a **local-only** companion that augments selected jobs with an LLM review via `src/lib/llm.ts` (auto-detects `claude`/`codex`/`gemini`/`opencode`, override `PUPILA_LLM`). Uses the local subscription — **not** an API key, so no per-token charges. Output: `data/ai-reviews.json`.
 
 Daily workflow:
 

README.md

@@ -95,16 +95,16 @@ The repo ships with neutral defaults. To make it yours:
 ### 1. Clone or fork
 
 ```bash
-gh repo fork FranRom/job-hunt --clone
-cd job-hunt
+gh repo fork FranRom/pupila --clone
+cd pupila
 pnpm install
 ```
 
 Or click "Fork" on GitHub, then `git clone <your-fork>`.
 
 ### 2. Generate your candidate brief (required)
 
-The brief at `config/candidate-brief.md` is the natural-language description of who you are, what you want, and what to avoid. **This step is mandatory** — `pnpm run dev` will refuse to start until the file exists. (Bypass with `JOB_HUNT_NO_BRIEF_CHECK=1` if you genuinely want raw aggregation with no AI review.)
+The brief at `config/candidate-brief.md` is the natural-language description of who you are, what you want, and what to avoid. **This step is mandatory** — `pnpm run dev` will refuse to start until the file exists. (Bypass with `PUPILA_NO_BRIEF_CHECK=1` if you genuinely want raw aggregation with no AI review.)
 
 **The file is gitignored** — it contains CV-derived personal information and should never be committed. setup-brief and the onboarding wizard write to the gitignored canonical path.
 
@@ -125,7 +125,7 @@ Or open the UI and use the Profile tab:
 pnpm run ui   # http://127.0.0.1:5173 → Profile tab → drop your CV
 ```
 
-The auto-detected provider order is `claude` → `codex` → `gemini` → `opencode` (whichever is on `PATH` first). Override with `JOB_HUNT_LLM=codex pnpm run setup-brief ...`. No API keys; uses your existing CLI subscription.
+The auto-detected provider order is `claude` → `codex` → `gemini` → `opencode` (whichever is on `PATH` first). Override with `PUPILA_LLM=codex pnpm run setup-brief ...`. No API keys; uses your existing CLI subscription.
 
 > **The two personalization layers, briefly:**
 > - `config/profile.json` (committed defaults) controls **what gets fetched + scored** (weights, keyword lists, tier-S slugs).
@@ -165,8 +165,8 @@ Two agents run on independent schedules so you can tune them separately:
 ./scripts/install-launchd.sh --review-time 09:00
 ./scripts/install-launchd.sh --no-review                  # aggregator only (no LLM CLI)
 ./scripts/install-launchd.sh --uninstall                  # remove both
-launchctl list | grep job-hunt                            # check status
-launchctl start dev.${USER}.job-hunt.aggregate            # trigger now
+launchctl list | grep pupila                               # check status
+launchctl start dev.${USER}.pupila.aggregate               # trigger now
 ```
 
 launchd's `StartCalendarInterval` catches up missed runs after wake — if your laptop was asleep at 7am, it runs once the lid opens.
@@ -548,7 +548,7 @@ The Jobs filter bar adds two unified-skip / unified-queue controls:
 
 [`src/ai-review.ts`](./src/ai-review.ts) is an **optional, local-only** companion that adds an LLM "second opinion" to selected jobs. Each job gets a structured review — summary, what they want, what they offer, red flags, and a verdict (`strong-match | match | weak-match | skip`) — so you can scan the day's matches in seconds instead of reading every posting.
 
-It shells out through [`src/lib/llm.ts`](./src/lib/llm.ts) to whichever local LLM CLI is available (`claude`, `codex`, `gemini`, or `opencode`; override with `JOB_HUNT_LLM=<provider>`). There are no project API keys and no per-token billing from this repo, but this cannot run in CI because the workflow runner is not authenticated as your local CLI user. Run it locally after the daily pipeline.
+It shells out through [`src/lib/llm.ts`](./src/lib/llm.ts) to whichever local LLM CLI is available (`claude`, `codex`, `gemini`, or `opencode`; override with `PUPILA_LLM=<provider>`). There are no project API keys and no per-token billing from this repo, but this cannot run in CI because the workflow runner is not authenticated as your local CLI user. Run it locally after the daily pipeline.
 
 ### One-time setup
 
@@ -733,7 +733,7 @@ Keyword arrays are joined with `|` and compiled into word-bounded, case-insensit
 <summary>Click to expand the full tree</summary>
 
 ```
-job-hunt/
+pupila/
 ├── .github/
 │   ├── workflows/
 │   │   └── check.yml          # PR/push: biome + typecheck + tests + build + audit
@@ -831,7 +831,7 @@ pnpm run clean:onboarding    # wipe only the onboarding state (preferences + bri
 pnpm run clean -- --all      # full fresh-clone reset (see "Reset to a clean slate" below)
 ```
 
-> **Heads up:** `pnpm run dev` refuses to start unless `config/candidate-brief.md` exists — set up your candidate brief first via `pnpm run setup-brief` or the UI Profile tab. Bypass with `JOB_HUNT_NO_BRIEF_CHECK=1` (or `--no-brief-check`) for raw aggregation without AI review.
+> **Heads up:** `pnpm run dev` refuses to start unless `config/candidate-brief.md` exists — set up your candidate brief first via `pnpm run setup-brief` or the UI Profile tab. Bypass with `PUPILA_NO_BRIEF_CHECK=1` (or `--no-brief-check`) for raw aggregation without AI review.
 
 The pre-commit hook runs `lint && typecheck` on every commit. To bypass it for an emergency commit: `SKIP_SIMPLE_GIT_HOOKS=1 git commit ...`.
 

package.json

@@ -1,5 +1,5 @@
 {
-  "name": "job-hunt",
+  "name": "pupila",
   "version": "0.1.0",
   "private": true,
   "type": "module",

scripts/install-cron.sh

@@ -19,8 +19,8 @@
 set -euo pipefail
 
 REPO_ROOT="$(cd "$(dirname "$0")/.." && pwd)"
-TAG_AGG="# job-hunt:aggregate:${REPO_ROOT}"
-TAG_REV="# job-hunt:review:${REPO_ROOT}"
+TAG_AGG="# pupila:aggregate:${REPO_ROOT}"
+TAG_REV="# pupila:review:${REPO_ROOT}"
 
 AGG_TIME="07:00"
 REV_TIME="07:15"
@@ -72,14 +72,14 @@ if [[ -z "$PNPM" ]]; then
   exit 1
 fi
 
-# Read current crontab (no error if empty), strip any existing job-hunt
+# Read current crontab (no error if empty), strip any existing pupila
 # entries for this repo so re-running is idempotent.
 CURRENT="$(crontab -l 2>/dev/null || true)"
 CLEANED="$(printf '%s\n' "$CURRENT" | grep -vF "$TAG_AGG" | grep -vF "$TAG_REV" || true)"
 
 if [[ "$UNINSTALL" == "1" ]]; then
   printf '%s\n' "$CLEANED" | crontab -
-  echo "✓ Removed job-hunt cron entries for $REPO_ROOT"
+  echo "✓ Removed pupila cron entries for $REPO_ROOT"
   exit 0
 fi
 

scripts/install-launchd.sh

@@ -1,10 +1,10 @@
 #!/usr/bin/env bash
 # Install two launchd agents on macOS:
 #
-#   1. dev.${USER}.job-hunt.aggregate — runs `pnpm run dev` (fetch + filter
+#   1. dev.${USER}.pupila.aggregate — runs `pnpm run dev` (fetch + filter
 #      + score + write data/jobs.json + JOBS.md + feed.xml). No LLM needed.
 #
-#   2. dev.${USER}.job-hunt.review    — runs `pnpm run ai-review` (per-job
+#   2. dev.${USER}.pupila.review    — runs `pnpm run ai-review` (per-job
 #      LLM verdict via your local CLI: claude / codex / gemini / opencode).
 #      Skipped via --no-review for users without an LLM CLI installed.
 #
@@ -20,7 +20,7 @@
 
 set -euo pipefail
 
-LABEL_BASE="dev.${USER}.job-hunt"
+LABEL_BASE="dev.${USER}.pupila"
 LAUNCH_DIR="$HOME/Library/LaunchAgents"
 AGG_LABEL="${LABEL_BASE}.aggregate"
 REV_LABEL="${LABEL_BASE}.review"
@@ -172,7 +172,7 @@ Trigger now:
   launchctl start ${REV_LABEL}
 
 Status:
-  launchctl list | grep job-hunt
+  launchctl list | grep pupila
 
 Uninstall:
   $0 --uninstall

scripts/install-mcp.sh

@@ -4,13 +4,13 @@
 # the client config(s).
 #
 # Usage:
-#   curl -sSf https://raw.githubusercontent.com/ogarciarevett/job-hunt/main/scripts/install-mcp.sh | bash
+#   curl -sSf https://raw.githubusercontent.com/FranRom/pupila/main/scripts/install-mcp.sh | bash
 #   # or, if you've cloned the repo already:
 #   bash scripts/install-mcp.sh
 #
 # Env overrides:
 #   PUPILA_HOME    - install location (default: $HOME/.pupila)
-#   PUPILA_REPO    - git URL (default: https://github.com/ogarciarevett/job-hunt.git)
+#   PUPILA_REPO    - git URL (default: https://github.com/FranRom/pupila.git)
 #   PUPILA_REF     - branch/tag/commit to checkout (default: main)
 #   PUPILA_DRY_RUN - if set to 1, print intended actions and exit
 #
@@ -25,7 +25,7 @@
 set -euo pipefail
 
 PUPILA_HOME="${PUPILA_HOME:-$HOME/.pupila}"
-PUPILA_REPO="${PUPILA_REPO:-https://github.com/ogarciarevett/job-hunt.git}"
+PUPILA_REPO="${PUPILA_REPO:-https://github.com/FranRom/pupila.git}"
 PUPILA_REF="${PUPILA_REF:-main}"
 PUPILA_DRY_RUN="${PUPILA_DRY_RUN:-0}"