Examples of Good Alerts

This page shows the shape of PolitiClaw's proactive output: what a well-formed rep-vote hit looks like, what silence looks like, what a weekly digest looks like.

About these examples

These examples are illustrative, not generated. The durable contracts live in:

• Generated Cron Jobs — exact schedules and payloads
• Generated Skills — the enforced alert-shape rules
• The skill markdown itself, read by the agent each time a cron fires

If an example here reads like a fictionalized version of the current output, that's the point: it's calibrated to the skill contract, not copy-pasted from a real run. Trust the skill files when the two drift.

A rep-vote watch hit

Emitted by politiclaw.rep_vote_watch when a bill on one of your tracked issues shows [new] or [changed] and crosses the confidence floor.

HR-1234 — Affordable Housing Accelerator Act [new]

Introduced 2026-04-18. Latest action: referred to House Financial Services.

Direction: appears to advance your support stance on affordable-housing. The bill text grants HUD authority to "waive local single-family-only zoning requirements where a state has declared a housing emergency" — which matches your stated preference for zoning-waiver expansion.

Counter-consideration from the tool: the same section preempts local zoning authority, which some civic groups argue undercuts neighborhood input.

Source: api.congress.gov (tier 1).

This alignment is a recorded claim, not a vote recommendation. Verify against the linked bill text before deciding.

What to notice:

• The [new] tag came from the plugin's change detection, not the agent.
• The direction is quoted from the bill's own text. A bill with no quotable directional signal renders as direction unclear instead.
• The alignment disclaimer is present verbatim; the skill forbids stripping it.

A quiet-window silent-ok

Emitted by any monitoring job when the delta is empty.

No new or materially changed items since last check (checked 14 bills, 3 upcoming events).

That's the whole message. No padding, no "here's what's still on the watch list," no "stay tuned." The monitoring skill is explicit: empty deltas get one brief confirmation line, never a padded digest.

A weekly summary

Emitted by politiclaw.weekly_summary on the 7-day cadence. Target length is 250–400 words, enforced by the summary skill.

Headline. Quiet week on federal climate bills; one committee markup moved on your oppose stance on public-lands-and-natural-resources.

Bills touching your declared stances.

• HR-2211 — Wetlands Regulatory Clarity Act · status: reported favorably from House Natural Resources, 23–19 · aligns against your oppose on public-lands-and-natural-resources (tier 1, api.congress.gov)
• S. 845 — Clean Energy Manufacturing Tax Credit Extension · status: introduced · aligns with your support on energy (tier 1, api.congress.gov)

Upcoming (next 7 days).

• House Natural Resources markup on HR-2211 — 2026-04-29, 10:00 ET
• House Ways and Means hearing on clean-energy credits — 2026-04-30, 14:00 ET

Things you might be surprised by.

HR-2211's committee report cites testimony from two wetlands scientists you would likely agree with arguing the bill's permit-streamlining language could be narrowed without gutting the underlying protections. Worth reading their framing before assuming the bill is purely rollback (source: House Natural Resources committee report, tier 1).

What PolitiClaw missed.

No source issues this week.

This summary is informational, not independent journalism. Verify against neutral sources before voting or contacting officials.

What to notice:

• The "things you might be surprised by" section is mandatory on any week with bill movement. If the week's delta is genuinely one-directional, the skill says so explicitly rather than fabricating opposition.
• Every numerical claim carries a source tier.
• The disclaimer is the exact verbatim line the skill requires.

An election-proximity ping

Emitted by politiclaw.election_proximity_alert only at 30, 14, 7, and 1 day before an election on your saved ballot.

Election in 14 days at 1234 Civic Center Dr, Springfield. Run politiclaw_election_brief for a full guide.

One line. Other days produce nothing.

A rep report excerpt

Emitted by politiclaw.rep_report on the 30-day cadence. The full digest covers every stored rep; this is the shape of a single entry.

Rep. Jane Doe (CA-12)

Recorded-vote alignment on your declared stances: 7 of 9 counted votes align; 2 against on tracked issues.

• Aligned: voted YES on HR-1234 (affordable-housing waivers; tier 1, api.congress.gov)
• Against: voted YES on HR-2211 (wetlands rollback; your oppose on public-lands-and-natural-resources)

Blind spots: 3 bills matched your stances but had no stance signal because they did not reach a recorded roll call.

This alignment is a recorded claim, not a vote recommendation. Verify against the linked bill text before deciding.

What to notice:

• The counted-votes denominator is explicit ("7 of 9 counted votes") — the skill forbids presenting partial coverage as a complete record.
• Blind spots are named inline, not buried in a footer.
• Links resolve to congress.gov tier-1 primary sources.

What a bad alert looks like (and why PolitiClaw won't produce one)

Three counter-examples the skills explicitly prevent:

A fabricated dissenting view

Things you might be surprised by. Critics of HR-1234 argue it could raise housing costs in suburban markets.

Missing: named source, tier tag, quotable basis. The monitoring skill requires the dissenting item to be tier 1–3 with a source link, and forbids LLM-authored counterpoints. If no real dissenting item exists in the delta, the skill says so explicitly.

An LLM-search vote tally

Rep. Doe has voted against wetlands protection 62% of the time this session.

Percentages without a tier-1 vote ledger are a tier-5 fabrication. The skill is explicit: no numerical claim gets attributed to LLM search. If the recorded-vote set doesn't support a number, the correct output is "number not verifiable from deterministic sources."

A padded empty week

Headline. A quiet but interesting week, with lots happening behind the scenes on your tracked issues...

Padding empty deltas trains the user to stop reading. The skill requires a single quiet-week line instead. Brevity is a trust mechanism, not a missing feature.

In the accountability loop

This page is the "what alerts look like" side of the loop. The rest:

• How PolitiClaw Holds Representatives Accountable — the full loop with the diagram, the dissenting-view rule, and source-tier discipline.
• See How My Reps Align — the operational entry point that scoring drives.
• Track Bills and Votes — the evidence side that monitoring jobs read from.