jobtriage 0.3.0__tar.gz
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- jobtriage-0.3.0/LICENSE +21 -0
- jobtriage-0.3.0/PKG-INFO +364 -0
- jobtriage-0.3.0/README.md +335 -0
- jobtriage-0.3.0/pyproject.toml +66 -0
- jobtriage-0.3.0/pyproject.toml.orig +62 -0
- jobtriage-0.3.0/src/jobtriage/__init__.py +10 -0
- jobtriage-0.3.0/src/jobtriage/__main__.py +7 -0
- jobtriage-0.3.0/src/jobtriage/assets/AGENTS.md +79 -0
- jobtriage-0.3.0/src/jobtriage/assets/gitignore +9 -0
- jobtriage-0.3.0/src/jobtriage/assets/jobtriage.toml +40 -0
- jobtriage-0.3.0/src/jobtriage/assets/letters/README.md +3 -0
- jobtriage-0.3.0/src/jobtriage/assets/presets/any.toml +10 -0
- jobtriage-0.3.0/src/jobtriage/assets/presets/backend-engineer.toml +52 -0
- jobtriage-0.3.0/src/jobtriage/assets/presets/data.toml +37 -0
- jobtriage-0.3.0/src/jobtriage/assets/presets/designer.toml +43 -0
- jobtriage-0.3.0/src/jobtriage/assets/presets/product-manager.toml +38 -0
- jobtriage-0.3.0/src/jobtriage/assets/presets/web-engineer.toml +53 -0
- jobtriage-0.3.0/src/jobtriage/assets/profile/DETAILS.md +32 -0
- jobtriage-0.3.0/src/jobtriage/assets/profile/README.md +7 -0
- jobtriage-0.3.0/src/jobtriage/assets/scrapers/README.md +9 -0
- jobtriage-0.3.0/src/jobtriage/assets/scrapers/_example.py +229 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/README.md +14 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/add-scraper/SKILL.md +143 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/add-scraper/references/contract.md +101 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/cover-letter/SKILL.md +107 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/cover-letter/references/example-letter.md +34 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/fill-application/SKILL.md +124 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/fill-application/references/ats-forms.md +52 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/jobtriage/SKILL.md +112 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/jobtriage/references/cli.md +212 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/jobtriage-db/SKILL.md +104 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/jobtriage-db/references/recipes.sql +107 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/jobtriage-db/references/schema.md +82 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/setup/SKILL.md +121 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/triage/SKILL.md +101 -0
- jobtriage-0.3.0/src/jobtriage/assets/skills/tune-filters/SKILL.md +142 -0
- jobtriage-0.3.0/src/jobtriage/assets/writing-style.md +38 -0
- jobtriage-0.3.0/src/jobtriage/attachments.py +54 -0
- jobtriage-0.3.0/src/jobtriage/browser/__init__.py +9 -0
- jobtriage-0.3.0/src/jobtriage/browser/find.py +177 -0
- jobtriage-0.3.0/src/jobtriage/browser/front.py +152 -0
- jobtriage-0.3.0/src/jobtriage/chrome.py +283 -0
- jobtriage-0.3.0/src/jobtriage/cli/__init__.py +9 -0
- jobtriage-0.3.0/src/jobtriage/cli/attach.py +72 -0
- jobtriage-0.3.0/src/jobtriage/cli/config.py +388 -0
- jobtriage-0.3.0/src/jobtriage/cli/db.py +216 -0
- jobtriage-0.3.0/src/jobtriage/cli/doctor.py +468 -0
- jobtriage-0.3.0/src/jobtriage/cli/export.py +94 -0
- jobtriage-0.3.0/src/jobtriage/cli/filter.py +114 -0
- jobtriage-0.3.0/src/jobtriage/cli/filters.py +67 -0
- jobtriage-0.3.0/src/jobtriage/cli/import_sheet.py +407 -0
- jobtriage-0.3.0/src/jobtriage/cli/init.py +337 -0
- jobtriage-0.3.0/src/jobtriage/cli/list.py +67 -0
- jobtriage-0.3.0/src/jobtriage/cli/main.py +230 -0
- jobtriage-0.3.0/src/jobtriage/cli/mcp.py +37 -0
- jobtriage-0.3.0/src/jobtriage/cli/note.py +37 -0
- jobtriage-0.3.0/src/jobtriage/cli/open.py +59 -0
- jobtriage-0.3.0/src/jobtriage/cli/output.py +43 -0
- jobtriage-0.3.0/src/jobtriage/cli/scrape.py +296 -0
- jobtriage-0.3.0/src/jobtriage/cli/show.py +148 -0
- jobtriage-0.3.0/src/jobtriage/cli/skills.py +230 -0
- jobtriage-0.3.0/src/jobtriage/cli/sources.py +947 -0
- jobtriage-0.3.0/src/jobtriage/cli/status.py +58 -0
- jobtriage-0.3.0/src/jobtriage/cli/tty.py +26 -0
- jobtriage-0.3.0/src/jobtriage/cli/tui.py +56 -0
- jobtriage-0.3.0/src/jobtriage/config.py +1127 -0
- jobtriage-0.3.0/src/jobtriage/db.py +334 -0
- jobtriage-0.3.0/src/jobtriage/geo.py +1258 -0
- jobtriage-0.3.0/src/jobtriage/logs.py +137 -0
- jobtriage-0.3.0/src/jobtriage/mcp.py +498 -0
- jobtriage-0.3.0/src/jobtriage/pipeline.py +582 -0
- jobtriage-0.3.0/src/jobtriage/plugin/__init__.py +74 -0
- jobtriage-0.3.0/src/jobtriage/plugin/dates.py +161 -0
- jobtriage-0.3.0/src/jobtriage/plugin/geo.py +42 -0
- jobtriage-0.3.0/src/jobtriage/plugin/http.py +322 -0
- jobtriage-0.3.0/src/jobtriage/plugin/parse.py +104 -0
- jobtriage-0.3.0/src/jobtriage/registry.py +539 -0
- jobtriage-0.3.0/src/jobtriage/relevance.py +276 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/__init__.py +19 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/bruntwork.py +230 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/dynamitejobs.py +208 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/euremotejobs.py +330 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/eurotechjobs.py +172 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/inclusivelyremote.py +212 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/jobgether.py +377 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/justremote.py +146 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/remocate.py +356 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/remotefrontendjobs.py +256 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/remoteok.py +208 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/remotesource.py +124 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/remotive.py +187 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/wellfound.py +244 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/weworkremotely.py +400 -0
- jobtriage-0.3.0/src/jobtriage/scrapers/workingnomads.py +105 -0
- jobtriage-0.3.0/src/jobtriage/status.py +126 -0
- jobtriage-0.3.0/src/jobtriage/store.py +712 -0
- jobtriage-0.3.0/src/jobtriage/tui/__init__.py +27 -0
- jobtriage-0.3.0/src/jobtriage/tui/app.py +912 -0
- jobtriage-0.3.0/src/jobtriage/tui/help.py +82 -0
- jobtriage-0.3.0/src/jobtriage/tui/keys.py +264 -0
- jobtriage-0.3.0/src/jobtriage/tui/widgets.py +438 -0
- jobtriage-0.3.0/src/jobtriage/workspace.py +470 -0
jobtriage-0.3.0/LICENSE
ADDED
|
@@ -0,0 +1,21 @@
|
|
|
1
|
+
MIT License
|
|
2
|
+
|
|
3
|
+
Copyright (c) 2026 Niyi Adeosun
|
|
4
|
+
|
|
5
|
+
Permission is hereby granted, free of charge, to any person obtaining a copy
|
|
6
|
+
of this software and associated documentation files (the "Software"), to deal
|
|
7
|
+
in the Software without restriction, including without limitation the rights
|
|
8
|
+
to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
|
|
9
|
+
copies of the Software, and to permit persons to whom the Software is
|
|
10
|
+
furnished to do so, subject to the following conditions:
|
|
11
|
+
|
|
12
|
+
The above copyright notice and this permission notice shall be included in all
|
|
13
|
+
copies or substantial portions of the Software.
|
|
14
|
+
|
|
15
|
+
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
|
|
16
|
+
IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
|
|
17
|
+
FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
|
|
18
|
+
AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
|
|
19
|
+
LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
|
|
20
|
+
OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
|
|
21
|
+
SOFTWARE.
|
jobtriage-0.3.0/PKG-INFO
ADDED
|
@@ -0,0 +1,364 @@
|
|
|
1
|
+
Metadata-Version: 2.4
|
|
2
|
+
Name: jobtriage
|
|
3
|
+
Version: 0.3.0
|
|
4
|
+
Summary: Scrape remote job boards, keep the relevant roles, remember them in SQLite, triage them from the keyboard.
|
|
5
|
+
Keywords: jobs,job-search,job-boards,scraping,cli,sqlite,tui
|
|
6
|
+
Author: Niyi Adeosun
|
|
7
|
+
Author-email: Niyi Adeosun <deniyi.adeosun@gmail.com>
|
|
8
|
+
License-Expression: MIT
|
|
9
|
+
License-File: LICENSE
|
|
10
|
+
Classifier: Development Status :: 4 - Beta
|
|
11
|
+
Classifier: Environment :: Console
|
|
12
|
+
Classifier: Intended Audience :: End Users/Desktop
|
|
13
|
+
Classifier: Operating System :: MacOS
|
|
14
|
+
Classifier: Operating System :: Microsoft :: Windows
|
|
15
|
+
Classifier: Operating System :: POSIX :: Linux
|
|
16
|
+
Classifier: Programming Language :: Python :: 3
|
|
17
|
+
Classifier: Programming Language :: Python :: 3.11
|
|
18
|
+
Classifier: Programming Language :: Python :: 3.12
|
|
19
|
+
Classifier: Programming Language :: Python :: 3.13
|
|
20
|
+
Classifier: Topic :: Utilities
|
|
21
|
+
Requires-Dist: mcp>=2.3.0
|
|
22
|
+
Requires-Dist: platformdirs>=4.3.0
|
|
23
|
+
Requires-Dist: textual>=8.2.8
|
|
24
|
+
Requires-Dist: tomlkit>=0.15.2
|
|
25
|
+
Requires-Python: >=3.11
|
|
26
|
+
Project-URL: Homepage, https://github.com/niyiao/jobtriage
|
|
27
|
+
Project-URL: Repository, https://github.com/niyiao/jobtriage
|
|
28
|
+
Description-Content-Type: text/markdown
|
|
29
|
+
|
|
30
|
+
# jobtriage
|
|
31
|
+
|
|
32
|
+
One CLI over fifteen remote job boards — and one place to keep what it
|
|
33
|
+
finds. jobtriage scrapes every enabled board in parallel, drops the
|
|
34
|
+
postings older than three weeks, keeps the titles that match your
|
|
35
|
+
filter, and remembers every job and its triage status in a local SQLite
|
|
36
|
+
database, so a later scrape only shows you what is new. A TUI sits on
|
|
37
|
+
top: newest first, one key to move a job through **new → to apply →
|
|
38
|
+
applied → skip**, and `enter` opens the posting in a Chrome window it
|
|
39
|
+
keeps for applying.
|
|
40
|
+
|
|
41
|
+
<!-- screenshot: TODO — the repository has no real TUI screenshot yet.
|
|
42
|
+
Take one of `jobtriage` on a populated database (a wide terminal
|
|
43
|
+
with the details pane open), save it under docs/, and embed it
|
|
44
|
+
here. Never commit a mock-up. -->
|
|
45
|
+
|
|
46
|
+
## Install
|
|
47
|
+
|
|
48
|
+
uv tool install jobtriage # or: pipx install jobtriage
|
|
49
|
+
jobtriage init # create the op folder (default ~/jobtriage)
|
|
50
|
+
jobtriage doctor # check the setup; every problem names its fix
|
|
51
|
+
|
|
52
|
+
Python 3.11 or newer. macOS and Linux are the tested platforms.
|
|
53
|
+
Windows is supported in the code — the browser finder, the launch, the
|
|
54
|
+
copy fallback for skill links and the CI matrix all cover it — but the
|
|
55
|
+
TUI has not been run in Windows Terminal yet. If the screen renders
|
|
56
|
+
badly there, a screenshot in an issue is the report that helps.
|
|
57
|
+
|
|
58
|
+
`jobtriage init` asks four small questions — where you can work from,
|
|
59
|
+
what kind of role, a title-filter preset, employment — then writes the
|
|
60
|
+
op folder from the shipped templates. Pressing Enter keeps each
|
|
61
|
+
default; `--yes` takes all four without asking, for scripts and agents.
|
|
62
|
+
Everything personal lives in that folder:
|
|
63
|
+
|
|
64
|
+
~/jobtriage/
|
|
65
|
+
jobtriage.toml config; its presence marks the folder
|
|
66
|
+
jobtriage.db the database: jobs, statuses, notes
|
|
67
|
+
profile/ RESUME.md, DETAILS.md, resume PDF
|
|
68
|
+
letters/ cover letters
|
|
69
|
+
scrapers/ user scraper plugins
|
|
70
|
+
skills/ agent skills
|
|
71
|
+
AGENTS.md agent instructions for this folder
|
|
72
|
+
CLAUDE.md symlink to AGENTS.md
|
|
73
|
+
.agents/skills symlink to skills/ (also .claude/skills,
|
|
74
|
+
.opencode/skills)
|
|
75
|
+
writing-style.md prose style for letters and content
|
|
76
|
+
.env secrets (API keys), optional
|
|
77
|
+
.browser-profile/ the Chrome profile for applying/research
|
|
78
|
+
.jobtriage/ tool state: version stamp, asset manifest,
|
|
79
|
+
database backups, logs, the TUI's remembered place
|
|
80
|
+
.gitignore so the folder can be git-tracked safely
|
|
81
|
+
|
|
82
|
+
The repository holds the code and nothing personal. The command finds
|
|
83
|
+
the folder by itself: from inside it, from the pointer `init` wrote, or
|
|
84
|
+
explicitly with `--workspace PATH` or `JOBTRIAGE_HOME`.
|
|
85
|
+
|
|
86
|
+
`init` is idempotent — safe to run twice, it never overwrites a file
|
|
87
|
+
you've edited. After a package update, `jobtriage init --upgrade`
|
|
88
|
+
refreshes the shipped templates: files you haven't touched are
|
|
89
|
+
replaced, edited files are left alone and the new version lands beside
|
|
90
|
+
them as `<file>.new`. If the package is newer than the folder, the
|
|
91
|
+
first run says so and points at that command.
|
|
92
|
+
|
|
93
|
+
## Quickstart
|
|
94
|
+
|
|
95
|
+
jobtriage scrape # the pipeline: every board in parallel,
|
|
96
|
+
# relevant new jobs into the database
|
|
97
|
+
jobtriage # the TUI over what landed
|
|
98
|
+
|
|
99
|
+
A default scrape asks each enabled board for its 50 newest jobs for
|
|
100
|
+
your configured locations, merges and dedupes them, drops anything
|
|
101
|
+
older than 21 days, keeps the titles the filter matches, and ingests
|
|
102
|
+
what remains in one transaction. The one-line summary
|
|
103
|
+
(`N relevant · x new · y already seen · total in the database`) goes to
|
|
104
|
+
stderr. A second run reports what it has already seen; nothing is
|
|
105
|
+
repeated.
|
|
106
|
+
|
|
107
|
+
`jobtriage scrape -s rs 20` runs one board with a different limit;
|
|
108
|
+
`--raw` prints the unfiltered listing and saves nothing;
|
|
109
|
+
`--explain` says, per board, what the age and title gates dropped and
|
|
110
|
+
which words did it.
|
|
111
|
+
|
|
112
|
+
### The TUI
|
|
113
|
+
|
|
114
|
+
Tabs follow the workflow: **new → to apply → applied → skip → all**,
|
|
115
|
+
opening on `tui.default_filter` (to apply) unless the last session
|
|
116
|
+
ended on another. `←`/`→` change tabs and never touch a status. The
|
|
117
|
+
tab, the search query and the selected job are saved on quit and
|
|
118
|
+
restored at the next start.
|
|
119
|
+
|
|
120
|
+
A status key writes the status the moment it is pressed and moves on
|
|
121
|
+
to the next job — triage is a stream of decisions. `u` undoes the last
|
|
122
|
+
one: status, tab and cursor, repeatable for the session. Writes go
|
|
123
|
+
straight to the database, and every change lands in the job's history
|
|
124
|
+
with the actor `tui`.
|
|
125
|
+
|
|
126
|
+
`/` searches title and company as you type; enter keeps the query, esc
|
|
127
|
+
clears it. `m` adds a one-line note, `y` copies the job's URL, `d`
|
|
128
|
+
shows the details pane (company, dates, locations, pay, source, notes,
|
|
129
|
+
history, attachments), and `?` shows every binding. `R` runs a scrape
|
|
130
|
+
in the background and reloads the list when it lands; `r` reloads
|
|
131
|
+
without scraping.
|
|
132
|
+
|
|
133
|
+
`enter` opens the job in the apply Chrome — the op folder's browser
|
|
134
|
+
profile, started if it isn't already up. Each job remembers its tab: a
|
|
135
|
+
second enter focuses that tab, and a new one opens only once the old
|
|
136
|
+
has been closed. Quitting the TUI leaves Chrome running. With no
|
|
137
|
+
Chrome installed, `enter` opens the job in the system's default
|
|
138
|
+
browser instead.
|
|
139
|
+
|
|
140
|
+
<!-- tui keys: generated from src/jobtriage/tui/keys.py — do not edit by hand -->
|
|
141
|
+
|
|
142
|
+
| Key | Action |
|
|
143
|
+
|---|---|
|
|
144
|
+
| `n` | set new and move on |
|
|
145
|
+
| `t` | set to apply and move on |
|
|
146
|
+
| `a` | set applied and move on |
|
|
147
|
+
| `s` | set skip and move on |
|
|
148
|
+
| `u` | undo the last status change |
|
|
149
|
+
| `j` / `↓` | move down one job |
|
|
150
|
+
| `k` / `↑` | move up one job |
|
|
151
|
+
| `g` | press twice for the first job |
|
|
152
|
+
| `G` | go to the last job |
|
|
153
|
+
| `^d` | move half a screen down |
|
|
154
|
+
| `^u` | move half a screen up |
|
|
155
|
+
| `pgdn` | move a page down |
|
|
156
|
+
| `pgup` | move a page up |
|
|
157
|
+
| `←` / `h` / `⇧tab` | previous tab |
|
|
158
|
+
| `→` / `l` / `tab` | next tab |
|
|
159
|
+
| `1` | jump to the new tab |
|
|
160
|
+
| `2` | jump to the to apply tab |
|
|
161
|
+
| `3` | jump to the applied tab |
|
|
162
|
+
| `4` | jump to the skip tab |
|
|
163
|
+
| `5` | jump to the all tab |
|
|
164
|
+
| `space` | space twice: applied (alias for a) |
|
|
165
|
+
| `S` | skip and move on (alias for s) |
|
|
166
|
+
| `enter` | open the job in Chrome |
|
|
167
|
+
| `d` | toggle the details pane |
|
|
168
|
+
| `/` | search title and company |
|
|
169
|
+
| `y` | copy the job URL |
|
|
170
|
+
| `m` | add a note to the job |
|
|
171
|
+
| `r` | reload from the database |
|
|
172
|
+
| `R` | run a scrape in the background |
|
|
173
|
+
| `?` | show this help |
|
|
174
|
+
| `q` / `^c` / `^q` | quit |
|
|
175
|
+
|
|
176
|
+
<!-- end tui keys -->
|
|
177
|
+
|
|
178
|
+
Keys are rebindable per op folder; the help overlay and the footer
|
|
179
|
+
follow the same table:
|
|
180
|
+
|
|
181
|
+
[tui.keys]
|
|
182
|
+
skipped = "x"
|
|
183
|
+
|
|
184
|
+
## Commands
|
|
185
|
+
|
|
186
|
+
| Command | What it does |
|
|
187
|
+
|---|---|
|
|
188
|
+
| `jobtriage` | the TUI (same as `jobtriage tui`) |
|
|
189
|
+
| `jobtriage scrape [LIMIT]` | run the boards; `-s NAME`, `--raw`, `--explain` |
|
|
190
|
+
| `jobtriage sources` | the boards: list, `info`, `new`, `test`, `enable`, `disable` |
|
|
191
|
+
| `jobtriage list [LIMIT]` | query the database; `--status`, `--since`, `--source`, `--search`, `--json` |
|
|
192
|
+
| `jobtriage show ID` | one job: every field, notes, history, attachments |
|
|
193
|
+
| `jobtriage open ID` | the posting in the apply browser |
|
|
194
|
+
| `jobtriage status ID SLUG` | set a job's status |
|
|
195
|
+
| `jobtriage note ID TEXT` | append a note |
|
|
196
|
+
| `jobtriage attach ID FILE` | record a file (a cover letter) against a job |
|
|
197
|
+
| `jobtriage export` | the rows as CSV, or `--format json` |
|
|
198
|
+
| `jobtriage config` | `show`, `edit`, `get`, `set`, `validate` |
|
|
199
|
+
| `jobtriage filter test TITLE` | why one title passes or not |
|
|
200
|
+
| `jobtriage db` | `path`, `shell`, `schema`, `import-sheet old.csv` |
|
|
201
|
+
| `jobtriage skills` | list the agent skills; `install --global --agent …` |
|
|
202
|
+
| `jobtriage mcp` | the MCP server over stdio |
|
|
203
|
+
| `jobtriage doctor [--offline]` | check the setup |
|
|
204
|
+
| `jobtriage where` | print the op folder's path |
|
|
205
|
+
| `jobtriage --version` | build, workspace path, database schema |
|
|
206
|
+
|
|
207
|
+
Every command takes `--help`; `jobtriage <cmd> --help` is the reference
|
|
208
|
+
for its flags. The config keys are in [docs/config.md](docs/config.md),
|
|
209
|
+
the boards and the plugin contract in
|
|
210
|
+
[docs/scrapers.md](docs/scrapers.md). `-v`/`--verbose` before the
|
|
211
|
+
command puts DEBUG detail on stderr. (`uv run jobtriage …` works from a
|
|
212
|
+
checkout; `python -m jobtriage` too.)
|
|
213
|
+
|
|
214
|
+
## Filtering
|
|
215
|
+
|
|
216
|
+
Two gates. **Age**: anything older than `search.max_age_days` (21 by
|
|
217
|
+
default) is dropped; a job with no date at all is assumed new. **Title**:
|
|
218
|
+
the `[filter]` table, starting from a shipped preset. A fresh folder
|
|
219
|
+
has no title filter (`preset = "any"`), so the board's category is the
|
|
220
|
+
only narrowing.
|
|
221
|
+
|
|
222
|
+
`jobtriage filter test "Senior React Native Engineer"` says why one
|
|
223
|
+
title passes or not. Every key and every preset is in
|
|
224
|
+
[docs/config.md](docs/config.md#filter); the reasoning behind the age
|
|
225
|
+
cutoff and the undated rule — and how the dedup keys work — is in
|
|
226
|
+
[docs/design.md](docs/design.md).
|
|
227
|
+
|
|
228
|
+
## The database
|
|
229
|
+
|
|
230
|
+
`jobtriage.db` in the op folder (gitignored by the shipped template) is
|
|
231
|
+
the single source of truth: every job the pipeline has seen, its
|
|
232
|
+
status, your notes, and a history row for every status change. `list`
|
|
233
|
+
and `export` read it without scraping; `show` prints one job in full;
|
|
234
|
+
`status`, `note` and `attach` are the writes.
|
|
235
|
+
|
|
236
|
+
Ad hoc questions go through sqlite3:
|
|
237
|
+
|
|
238
|
+
jobtriage db shell 'select title, company, url from jobs order by published desc limit 10'
|
|
239
|
+
jobtriage db schema # the tables and columns, with comments
|
|
240
|
+
|
|
241
|
+
Coming from an old Google Sheet: `jobtriage db import-sheet old.csv`
|
|
242
|
+
(File → Download → CSV) moves its statuses over, matching by url first
|
|
243
|
+
and then by folded title and company; `--dry-run` first. It is a
|
|
244
|
+
one-time bridge, not part of the daily workflow. The dedup key and the
|
|
245
|
+
column list are in
|
|
246
|
+
[docs/design.md](docs/design.md#dedup-keys-and-the-jobs-table).
|
|
247
|
+
|
|
248
|
+
## Config
|
|
249
|
+
|
|
250
|
+
Everything you might change is `jobtriage.toml` in the op folder: where
|
|
251
|
+
you can work from, families, age and the per-board limit, the title
|
|
252
|
+
filter, disabled boards and per-board overrides, the browser, the
|
|
253
|
+
scraper's User-Agent, the TUI's startup tab and key bindings. Every key
|
|
254
|
+
ships commented out, so a fresh folder runs on the defaults until you
|
|
255
|
+
fill something in.
|
|
256
|
+
|
|
257
|
+
`jobtriage config show` prints every effective value with the layer
|
|
258
|
+
that set it, `config set` edits the file without losing its comments,
|
|
259
|
+
and `config validate` names a typo with a suggestion. The reference is
|
|
260
|
+
[docs/config.md](docs/config.md).
|
|
261
|
+
|
|
262
|
+
## Agents and skills
|
|
263
|
+
|
|
264
|
+
`init` ships eight agent skills in `<op>/skills/`: an umbrella
|
|
265
|
+
`jobtriage` page, the database and triage workflows, cover letters,
|
|
266
|
+
applications, scrapers, filter tuning and setup. It links them from
|
|
267
|
+
`.agents/skills`, `.claude/skills` and `.opencode/skills` (and
|
|
268
|
+
`CLAUDE.md` to `AGENTS.md`), so an agent opened in the folder finds
|
|
269
|
+
them by itself.
|
|
270
|
+
|
|
271
|
+
jobtriage skills # what ships and what's installed
|
|
272
|
+
jobtriage skills install --global --agent all # copy them for the agents
|
|
273
|
+
jobtriage mcp # a stdio MCP server instead
|
|
274
|
+
|
|
275
|
+
`skills install --global` copies them into `~/.claude/skills`,
|
|
276
|
+
`~/.codex/skills` and `~/.config/opencode/skills` for use outside the
|
|
277
|
+
folder. `jobtriage mcp` is for clients without skill support: it
|
|
278
|
+
exposes typed reads and writes (`list_jobs`, `get_job`, `set_status`,
|
|
279
|
+
`add_note`, `attach`, `scrape`, `filter_test`) and never submits an
|
|
280
|
+
application. Triggers, discovery folders, the tool list and client
|
|
281
|
+
config: [docs/skills.md](docs/skills.md).
|
|
282
|
+
|
|
283
|
+
## Plugins
|
|
284
|
+
|
|
285
|
+
Fifteen boards ship in the package. Any Python file in
|
|
286
|
+
`<op>/scrapers/` is also a board — no registration, no fork, no
|
|
287
|
+
reinstall: `jobtriage sources new myboard` writes one from the shipped
|
|
288
|
+
template, `sources test myboard` runs it with limit 5 and names any
|
|
289
|
+
contract problem, and `sources` shows every board and its health. The
|
|
290
|
+
plugin contract, the worked example, the research playbook and the
|
|
291
|
+
notes for every board are in [docs/scrapers.md](docs/scrapers.md).
|
|
292
|
+
|
|
293
|
+
## doctor
|
|
294
|
+
|
|
295
|
+
`doctor` is the first thing to run when something is off. One line per
|
|
296
|
+
check — the workspace it found, the config, the database schema, the
|
|
297
|
+
browser and its CDP port, every plugin, the shipped skills, the profile
|
|
298
|
+
files — and a `problem` line carries the command that fixes it. It exits
|
|
299
|
+
1 when anything is broken. Its only network use is one `HEAD` request
|
|
300
|
+
per enabled board to the board's homepage. `--offline` runs the local
|
|
301
|
+
checks only — workspace, config, database, plugins, skills, profile —
|
|
302
|
+
and skips the three that depend on the machine: the browser, its CDP
|
|
303
|
+
port and the boards. That is the mode CI uses, and the right one for a
|
|
304
|
+
headless box or a support request that starts with "is this install
|
|
305
|
+
sound?".
|
|
306
|
+
|
|
307
|
+
## Logging
|
|
308
|
+
|
|
309
|
+
Every command also appends to `.jobtriage/logs/jobtriage.log`: DEBUG
|
|
310
|
+
detail, rotated at about a megabyte with three older copies kept.
|
|
311
|
+
`-v`/`--verbose` before the command puts the same detail on stderr:
|
|
312
|
+
|
|
313
|
+
jobtriage -v scrape # requests, board timings, every warning
|
|
314
|
+
|
|
315
|
+
## Responsible use
|
|
316
|
+
|
|
317
|
+
jobtriage is one person's reader for boards that publish jobs publicly;
|
|
318
|
+
it is not a crawler farm.
|
|
319
|
+
|
|
320
|
+
- A default run asks each enabled board for its 50 newest postings, one
|
|
321
|
+
pass. Boards that paginate stay within their declared `MAX_PAGES`
|
|
322
|
+
cap, and `POLITE_DELAY` puts a minimum gap between a board's requests
|
|
323
|
+
(`src/jobtriage/plugin/http.py` enforces it for the whole run).
|
|
324
|
+
- Each built-in declares `TERMS` — what the board asks of a scraper:
|
|
325
|
+
a link back, a credit, a request rate. `jobtriage sources info NAME`
|
|
326
|
+
prints it. The URLs stored are the canonical ones several boards'
|
|
327
|
+
terms require (remotefrontendjobs' `/{id}` page, weworkremotely's own
|
|
328
|
+
links), and remoteok's "Remote OK" credit is carried with its link.
|
|
329
|
+
- `euremotejobs` ships disabled because its terms ban automated
|
|
330
|
+
collection outright. `jobtriage sources enable euremotejobs` is how
|
|
331
|
+
you accept them after reading `sources info euremotejobs`.
|
|
332
|
+
- Requests send a desktop-browser User-Agent by default, because several
|
|
333
|
+
boards refuse a generic one. If you would rather identify the tool,
|
|
334
|
+
set `[network] user_agent` in `jobtriage.toml` (see docs/config.md);
|
|
335
|
+
a board that demands its own string still gets it.
|
|
336
|
+
- Run it when you will read the results. There is no scheduler and no
|
|
337
|
+
queue — a scrape is one pass you started — and the only retry is a
|
|
338
|
+
single reconnect after a dropped connection.
|
|
339
|
+
|
|
340
|
+
## Privacy
|
|
341
|
+
|
|
342
|
+
No telemetry, no account, no server of ours. The tool reads and writes
|
|
343
|
+
only the op folder, and its network traffic goes to two places: the job
|
|
344
|
+
boards you enabled, and any agent or API you choose to run yourself.
|
|
345
|
+
Delete the op folder and everything jobtriage knows is gone; the only
|
|
346
|
+
other file it writes is the one-line pointer in your platform config
|
|
347
|
+
directory that says where that folder is.
|
|
348
|
+
|
|
349
|
+
## Documentation and contributing
|
|
350
|
+
|
|
351
|
+
- [docs/config.md](docs/config.md) — every config key, the location
|
|
352
|
+
vocabulary, the presets
|
|
353
|
+
- [docs/scrapers.md](docs/scrapers.md) — the plugin contract, discovery,
|
|
354
|
+
the board notes
|
|
355
|
+
- [docs/skills.md](docs/skills.md) — the eight skills and the MCP server
|
|
356
|
+
- [docs/design.md](docs/design.md) — why: the cache keys, the one
|
|
357
|
+
transaction, the location model, the registry tiers, the version tracks
|
|
358
|
+
- [docs/research.md](docs/research.md) — the investigations that chose
|
|
359
|
+
the boards
|
|
360
|
+
- [CONTRIBUTING.md](CONTRIBUTING.md) — dev setup, tests, how to add a
|
|
361
|
+
board, the commit convention
|
|
362
|
+
- [CHANGELOG.md](CHANGELOG.md) — what changed, release by release
|
|
363
|
+
- [SECURITY.md](SECURITY.md) — the trust model and how to report a
|
|
364
|
+
problem
|