Agent Skills: FreshRSS

FreshRSS

My self-hosted RSS reader. Use this skill for anything RSS-related: browsing articles, searching content, managing feeds, finding recommendations, answering questions about what I read.

Environment

Variables from sops-nix secrets:

• FRESHRSS_URL - Base URL of FreshRSS instance
• FRESHRSS_API_USER - API username
• FRESHRSS_API_PASSWORD - API password

Authentication

Required before all API calls:

AUTH_TOKEN=$(curl -s -X POST "$FRESHRSS_URL/api/greader.php/accounts/ClientLogin" \
  -d "Email=$FRESHRSS_API_USER" \
  -d "Passwd=$FRESHRSS_API_PASSWORD" | grep -oP 'Auth=\K.*')

Reading & Searching Articles

Get recent items (all feeds)

curl -s "$FRESHRSS_URL/api/greader.php/reader/api/0/stream/contents/reading-list?output=json&n=20" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" | jq '.items[] | {title, published: .published, origin: .origin.title, summary: .summary.content[:200], href: .alternate[0].href}'

Get unread items

curl -s "$FRESHRSS_URL/api/greader.php/reader/api/0/stream/contents/reading-list?output=json&n=50&xt=user/-/state/com.google/read" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" | jq '.items[] | {title, origin: .origin.title, href: .alternate[0].href}'

Get items from a specific feed

curl -s "$FRESHRSS_URL/api/greader.php/reader/api/0/stream/contents/feed%2fFEED_URL?output=json&n=20" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" | jq '.items[] | {title, published: .published, summary: .summary.content[:200]}'

Get items from a category

curl -s "$FRESHRSS_URL/api/greader.php/reader/api/0/stream/contents/user%2f-%2flabel%2fCATEGORY_NAME?output=json&n=20" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" | jq '.items[] | {title, origin: .origin.title, href: .alternate[0].href}'

Get starred items

curl -s "$FRESHRSS_URL/api/greader.php/reader/api/0/stream/contents/user%2f-%2fstate%2fcom.google%2fstarred?output=json&n=50" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" | jq '.items[] | {title, origin: .origin.title, href: .alternate[0].href}'

Search articles by keyword

FreshRSS GReader API does not have a native search endpoint. To search, fetch a large batch of items and filter client-side with jq:

curl -s "$FRESHRSS_URL/api/greader.php/reader/api/0/stream/contents/reading-list?output=json&n=200" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" | jq '[.items[] | select(.title | test("KEYWORD"; "i")) | {title, origin: .origin.title, href: .alternate[0].href}]'

For deeper search, also check summary content:

curl -s "$FRESHRSS_URL/api/greader.php/reader/api/0/stream/contents/reading-list?output=json&n=500" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" | jq '[.items[] | select((.title // "" | test("KEYWORD"; "i")) or (.summary.content // "" | test("KEYWORD"; "i"))) | {title, origin: .origin.title, href: .alternate[0].href}]'

Use &c=TIMESTAMP (unix epoch) to paginate through older items.

Get item stream IDs (lightweight, for pagination)

curl -s "$FRESHRSS_URL/api/greader.php/reader/api/0/stream/items/ids?output=json&n=100&s=reading-list" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" | jq '.itemRefs[].id'

Feed Management

List all feeds

curl -s "$FRESHRSS_URL/api/greader.php/reader/api/0/subscription/list?output=json" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" | jq '.subscriptions[] | {title, id, url: .url, categories: [.categories[].label]}'

List categories

curl -s "$FRESHRSS_URL/api/greader.php/reader/api/0/tag/list?output=json" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" | jq '.tags[].id' | grep label

Unread count (per feed)

curl -s "$FRESHRSS_URL/api/greader.php/reader/api/0/unread-count?output=json" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" | jq '.unreadcounts[] | {id, count}'

Add a feed

RESULT=$(curl -s -X POST "$FRESHRSS_URL/api/greader.php/reader/api/0/subscription/quickadd" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" \
  -d "quickadd=https://example.com/feed.xml")
FEED_ID=$(echo "$RESULT" | jq -r '.streamId')
curl -s -X POST "$FRESHRSS_URL/api/greader.php/reader/api/0/subscription/edit" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" \
  -d "ac=edit" \
  -d "s=$FEED_ID" \
  -d "a=user/-/label/CATEGORY_NAME"

Move feed between categories

curl -s -X POST "$FRESHRSS_URL/api/greader.php/reader/api/0/subscription/edit" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" \
  -d "ac=edit" \
  -d "s=feed/ID" \
  -d "a=user/-/label/NewCategory" \
  -d "r=user/-/label/OldCategory"

Remove a feed

curl -s -X POST "$FRESHRSS_URL/api/greader.php/reader/api/0/subscription/edit" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" \
  -d "ac=unsubscribe" \
  -d "s=feed/FEED_ID"

Mark as read / star

# Mark item read
curl -s -X POST "$FRESHRSS_URL/api/greader.php/reader/api/0/edit-tag" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" \
  -d "i=ITEM_ID" \
  -d "a=user/-/state/com.google/read"

# Star item
curl -s -X POST "$FRESHRSS_URL/api/greader.php/reader/api/0/edit-tag" \
  -H "Authorization: GoogleLogin auth=$AUTH_TOKEN" \
  -d "i=ITEM_ID" \
  -d "a=user/-/state/com.google/starred"

Recommendations

When asked for feed recommendations:

1. List all current subscriptions with categories
2. Identify themes and interests from the feed titles and categories
3. Use WebSearch to find feeds in similar topic areas that the user does not already subscribe to
4. Present recommendations grouped by interest area, with feed URL and a brief description of what the feed covers

Workflow: add and categorize a feed

1. Authenticate and fetch the existing categories.
2. Fetch the feed URL and sample its title, description, and recent posts.
3. Prefer the best-fitting existing category. Propose a short category matching the current naming style only when none fits.
4. Confirm the category unless the match is obvious.
5. Add the feed using subscription/quickadd, then assign the category using subscription/edit.
6. Verify the resulting subscription and report its title, URL, and category.

Workflow: subscribe to an email newsletter

1. Fetch the newsletter's site to understand its subject and choose a title.

2. Create a Kill the Newsletter inbox:

   FEED_SLUG=$(curl -s -X POST https://kill-the-newsletter.com/feeds \
     -H "CSRF-Protection: true" \
     -d "title=NEWSLETTER_TITLE" \
     -D - -o /dev/null | grep -oP 'location:.*feeds/\K[^\s]+')
   

3. Add https://kill-the-newsletter.com/feeds/${FEED_SLUG}.xml using the feed workflow above.

4. Report ${FEED_SLUG}@kill-the-newsletter.com, the private Atom feed URL, its FreshRSS category, and that the email address must be entered on the newsletter signup page. Confirmation messages will appear in FreshRSS.

5. Never share the Kill the Newsletter feed URL publicly.

Workflow: organize uncategorized feeds

1. Fetch existing categories and subscriptions with no category or only the Uncategorized category.
2. Sample each feed's site or RSS URL. Infer an existing category where possible; skip feeds with insufficient evidence rather than guessing.
3. Present one batch table containing each feed, proposed category, and brief rationale. Let the user approve, reject, or override the proposals.
4. For approved feeds, add the selected category with subscription/edit and remove user/-/label/Uncategorized.
5. Report categorized, skipped, and still-uncategorized feeds, including any newly created categories.

Notes

• Token is session-based, get a fresh one each invocation
• Feed IDs use format feed/https://example.com/rss
• The Fever API (/api/fever.php) is read-only, use GReader API for mutations
• URL-encode slashes in stream IDs when used in URL paths (%2f)
• n= controls batch size, c= is a continuation token (unix timestamp) for pagination
• To read full article content, use WebFetch on the article's href