Html Report

Generates a self-contained HTML dashboard (stat cards, Chart.js charts, filterable table) from the same SQLite tracker database used by application-manager — opens directly in a browser, no server required

HTML Report — Application Tracker Dashboard

When to Use This Skill

Trigger when the user asks to:

• Generate, build, or refresh a job search dashboard ("build me a dashboard", "generate the html report", "/html-report")
• Visualize their application pipeline ("show me a chart of my applications", "visualize my job search")
• Export a shareable, offline view of their tracker ("give me something I can open in a browser", "make a report I can send")
• Re-run the dashboard after syncing new applications ("refresh the dashboard", "update the report with my latest applications")

Keywords: html report, dashboard, visualize applications, application chart, job search report, funnel chart, /html-report

This skill is read-only: it never modifies the tracker database. Use the job-application-manager skill (or swelist tracker add/update) to change application status — this skill only renders what's already there.

Requires the sqlite tracker backend. If profile.md has Tracker backend: notion, this skill has nothing to read — see Edge Cases.

---

Setup

Always read skills/profile.md first. It's the same file job-application-manager reads, so both skills stay pointed at the same database.

Resolve before continuing:

1. Tracker backend — Integrations > Tracker Backend
   • Must be sqlite. If notion, tell the user this report needs the SQLite backend and stop (see Edge Cases) rather than fabricating data.
2. SQLite DB path — Integrations > Tracker Backend > SQLite DB path
   • Default: ~/.offerplus/applications.db
   • If the file doesn't exist: tell the user to run swelist tracker init (or sync via job-application-manager first) — do not generate an empty report silently.
3. Output path — argument if given, else reports/application-dashboard.html. Create the reports/ directory if it doesn't exist.

---

Config: Argument Parsing

Input	Behavior
No argument	Output to reports/application-dashboard.html
A path argument (e.g. /html-report ~/Desktop/report.html)	Use that path instead
--open flag	After writing, tell the user to open the file — the skill cannot launch a browser itself

---

Config: SQLite Schema (read-only)

Same table job-application-manager writes to. See that skill's "Config: SQLite Schema" section for the authoritative DDL and migration history. This skill never runs ALTER TABLE — it just adapts to whichever columns exist.

Full (v0.5.0) schema:

sql

CREATE TABLE applications (
  name           TEXT PRIMARY KEY,   -- "Company — Role Title"
  status         TEXT NOT NULL,
  job_id         TEXT,
  company        TEXT,
  applied_on     TEXT,               -- ISO date YYYY-MM-DD
  last_touch     TEXT,               -- ISO date of most recent email
  interview_date TEXT,               -- ISO date of next scheduled interview
  next_action    TEXT,               -- Waiting | Prep | Follow up | Send availability
  link           TEXT,               -- job posting or portal URL
  tags           TEXT,               -- comma-separated e.g. "Referral,Remote"
  notes          TEXT,
  updated_at     TEXT DEFAULT (datetime('now'))
);

Older (v0.2.x) schema only has name, status, job_id, applied_on, notes, updated_at — that's also exactly what swelist tracker export returns today. Check which columns are present before building the SELECT:

bash

sqlite3 <DB_PATH> "PRAGMA table_info(applications);"

If company, last_touch, interview_date, next_action, link, or tags are missing, treat them as empty for every row and omit any chart/table section that depends entirely on them (see Layout, Edge Cases).

---

Config: Status Normalisation

Map every row's status value to one of five canonical buckets before computing stats. Comparisons are case-insensitive. Covers both the current pipeline statuses and the older v0.2.x values, so the report works regardless of which schema generation wrote the row.

Status value	Canonical bucket	Colour
Applied / Received, OA, Recruiter screen, Not started	Active	#3b82f6
Interviewing, In progress	Interview	#f59e0b
Offer	Offer	#8b5cf6
Hired, Done	Hired	#22c55e
Rejected, Withdrawn	Rejected/Closed	#ef4444
any other/unrecognised value	Active (fallback) — note in the report footer as "N rows had an unmapped status"	#3b82f6

In progress (v0.2.x) can't be distinguished from a true "Interviewing" stage — bucket it as Interview and note in the footer that granular staging requires the v0.5.0 schema.

---

Workflow

text

INPUT: output_path (optional, default "reports/application-dashboard.html")
       open_flag (optional)

STEP 0  Load profile + resolve DB
  Read skills/profile.md → tracker_backend, DB_PATH
  IF tracker_backend != "sqlite": stop, explain (see Edge Cases)
  IF DB_PATH file does not exist: stop, tell user to run `swelist tracker init`

STEP 1  Parse arguments
  IF path argument given: output_path = that path
  ELSE: output_path = "reports/application-dashboard.html"
  mkdir -p dirname(output_path)

STEP 2  Collect data
  columns = sqlite3 DB_PATH "PRAGMA table_info(applications);"
  select_cols = ["name","status","job_id","applied_on","notes","updated_at"]
                + any of ["company","last_touch","interview_date","next_action","link","tags"]
                  that exist in columns
  rows = sqlite3 -json DB_PATH
    "SELECT {', '.join(select_cols)} FROM applications ORDER BY applied_on DESC, name ASC"
  FOR each row:
    row.company_display = row.company OR extract_company_from_name(row.name)  # part before " — "
    row.role_display    = extract_role_from_name(row.name)                    # part after " — "
    row.bucket           = normalise_status(row.status)   # Config: Status Normalisation
    row.tag_list          = split(row.tags, ",") if present else []

STEP 3  Compute summary stats
  total              = len(rows)
  by_bucket           = count rows grouped by row.bucket
  by_tag              = count occurrences of each tag across row.tag_list (flatten)
  by_month            = count rows grouped by YYYY-MM of row.applied_on (blank → "Unknown")
  resolved            = rows where bucket != "Active"
  past_screen         = rows where bucket IN ("Interview", "Offer", "Hired")
  funnel_rate_pct     = round(100 * len(past_screen) / total) if total else 0
  rejection_rate_pct  = round(100 * count(bucket == "Rejected/Closed") / len(resolved)) if resolved else 0
  funnel_stages       = { Applied: total,
                          Interview: count(bucket IN (Interview, Offer, Hired)),
                          Offer: count(bucket IN (Offer, Hired)),
                          Hired: count(bucket == Hired) }

STEP 4  Generate HTML (see Layout, Design spec, and Charts below)
  Build one self-contained HTML string:
    - inline <style> block (palette, cards, charts grid, table)
    - stat cards row
    - two 2-column chart rows (doughnut, tags bar / by-month bar, funnel bar)
    - filterable table (search input + status/company dropdowns, client-side JS)
    - footer: "Generated by Claude Code · swelist · {ISO date}"
  Charts render via Chart.js loaded from https://cdn.jsdelivr.net/npm/chart.js
    (only external dependency; wrap chart init in try/catch so a failed CDN
    fetch degrades to the stat cards instead of breaking the page)

STEP 5  Write and confirm
  Write the complete HTML to output_path (Write tool — overwrite if it exists)
  IF open_flag: tell the user to open the file themselves (cannot launch a browser)
  Print the Report Format summary below

---

Layout

text

┌─────────────────────────────────────────────┐
│  🔍 Job Search Dashboard    Generated: DATE  │
├──────┬──────┬──────┬──────┬──────────────────┤
│Active│Inter-│Offer │Hired │Rejected/Closed   │  ← stat cards
│  N   │view N│  N   │  N   │       N          │
├──────┴──────┴──────┴──────┴──────────────────┤
│  Status breakdown (doughnut) │ By tag (bar)   │  ← charts row
├─────────────────────────────────────────────  ┤
│  By month (bar)    │  Funnel (horizontal bar) │  ← charts row
├──────────────────────────────────────────────  ┤
│  Applications  [Status ▾] [Company ▾] [🔍 ...]│  ← table with filters
│  date │ company │ role │ status │ next action │ ...│
│  ...                                          │
└───────────────────────────────────────────────┘

Design Spec

• Colour palette: CSS custom properties, using the status colours in Config: Status Normalisation above.
• Font: system-ui stack, no web fonts.
• Stat cards: white background, subtle shadow, large bold number, label below, left border in the status colour.
• Charts: 2-column grid on wide screens, stacked on narrow (<768px).
• Table:
  • Alternating row shading.
  • Status column uses a coloured pill/badge (Config: Status Normalisation colours).
  • link column renders as a hyperlink (target="_blank" rel="noopener") if present; otherwise —.
  • Empty cells render as —.
  • Client-side filter: a text input filters rows across company + role; status and company dropdowns filter independently; all three combine with AND.
  • Rows sort newest-first by default (applied_on descending, then company A→Z).
• Responsive: usable at 900px+; not required to be fully responsive below that.
• Footer: Generated by Claude Code · swelist · {ISO date}.

Charts (Chart.js)

1. Status doughnut — one slice per status bucket, palette colours above.
2. By tag bar (horizontal) — count per inferred tag (Referral, Remote, Urgent, Visa — see job-application-manager's Tag Inference config), sorted descending. Omit entirely if no row has a tags value.
3. By month bar — application count per YYYY-MM of applied_on, oldest to newest.
4. Application funnel (horizontal bar) — Applied → Interview → Offer → Hired, each bar = count of rows reaching at least that stage.

All chart <canvas> elements get an aria-label. Wrap each chart in <div class="chart-card"><h3>Title</h3><canvas ...></canvas></div>.

Table Columns

Date (applied_on) · Company · Role · Status · Next action · Tags · Interview date · Notes (truncate to 80 chars, full text in a title tooltip) · Link (hyperlink or —)

Omit any column whose value is empty across every row, and omit next_action / interview_date / tags / link entirely if the column doesn't exist in the DB (v0.2.x schema).

---

File I/O Reference

Checking available columns

bash

sqlite3 ~/.offerplus/applications.db "PRAGMA table_info(applications);"

Reading tracker rows (full v0.5.0 schema)

bash

sqlite3 -json ~/.offerplus/applications.db \
  "SELECT name, status, job_id, company, applied_on, last_touch, interview_date,
          next_action, link, tags, notes, updated_at
   FROM applications ORDER BY applied_on DESC, name ASC;"

Reading tracker rows (older v0.2.x schema fallback)

bash

sqlite3 -json ~/.offerplus/applications.db \
  "SELECT name, status, job_id, applied_on, notes, updated_at
   FROM applications ORDER BY applied_on DESC, name ASC;"

Writing the report

text

Write(file_path=output_path, content=full_html_string)

---

Report Format

text

Dashboard generated: <output path>

Open it in any browser — no server needed.

Summary:
- Total applications: N
- Active: N · Interview: N · Offer: N · Hired: N · Rejected/Closed: N
- Funnel: N% progressed past resume screen

Re-run /html-report any time after syncing via application-manager
(or `swelist tracker add/update`) to refresh the dashboard.

(Omit the funnel line if total == 0; print "No applications tracked yet" instead of the stat breakdown.)

---

Design Principles

• Self-contained. One file, opens offline (Chart.js needs a CDN fetch once; everything else is local).
• Data-only. This skill reads and renders; it never writes to the tracker database.
• Idempotent. Re-running overwrites the previous report at the same path — no accumulation.
• Single source of truth. Reads the exact same SQLite database job-application-manager writes to — no separate export/CSV step, so the dashboard is never stale relative to a sync that already happened.
• Graceful on sparse data. With only a few rows, charts still render correctly for small N; the table remains the primary value. Never suppress charts just because N is small.
• No fabrication. Every number in the report comes directly from the tracker database. Never infer or estimate missing fields.

---

Edge Cases

Situation	Handling
Tracker backend in profile.md is notion, not sqlite	Stop and tell the user this skill currently only reads the SQLite tracker; offer to switch tracker_backend to sqlite in profile.md, or point at an explicit --db path if they maintain a separate local copy
SQLite DB file does not exist at the resolved path	Tell the user to run swelist tracker init (or do an initial sync via job-application-manager); do not generate an empty report
DB exists but applications table has zero rows	Still generate the HTML shell; show "No applications tracked yet" instead of stat cards/charts
DB uses the older v0.2.x schema (no company/tags/etc.)	Build the SELECT from only the columns that exist (Step 2); omit dependent chart/table columns entirely rather than erroring
A row's status value isn't in the normalisation table	Fall back to Active; count it in an "N rows had an unmapped status" footer note
applied_on is null or blank	Group under "Unknown" in the by-month chart; sort such rows last in the table
No row has a tags value	Omit the "By tag" chart entirely rather than rendering an empty one
link value doesn't start with http	Render as plain text, not a hyperlink
A table column (e.g. next_action) is empty for every row	Omit that column from the rendered table entirely
Output path's parent directory doesn't exist	Create it (mkdir -p) before writing
Re-running with the same output path	Overwrite; never append or version the filename
User is offline when the report is later opened	Chart.js CDN fetch fails; charts area shows nothing but stat cards and table still render — never let a failed script tag blank the whole page
--open flag passed	Tell the user the file is ready and where it is; the skill cannot launch a browser itself