Proposed work: harden source validation in live bounty checker scripts

[SueJianjian]

Summary

scripts/check_bounty_issue_states.py and scripts/check_live_bounty_closing_refs.py each use a required mutually-exclusive --input/--repo source selection, but both still branch on string truthiness at runtime:

• check_bounty_issue_states.py
  • data = _load_input(args.input) if args.input else load_live_data(args.repo, args.api_host)
  • if args.fix:
  • if args.input:
• check_live_bounty_closing_refs.py
  • data = _load_input(args.input) if args.input else load_live_data(args.repo, args.api_host, args.state, args.pr)

That means empty-string or whitespace-only source values can satisfy argparse's required source group, then still select the wrong runtime path or defer validation until a raw traceback / gh error.

Why it is useful

These two scripts are public bounty hygiene checks used around live MergeWork issue/PR maintenance:

• check_bounty_issue_states.py verifies that open public bounty rows still point at open GitHub issues, and can optionally --fix by reopening issues
• check_live_bounty_closing_refs.py checks PR bodies for Closes #... / similar references against still-open bounty issues

Malformed source arguments in those tools should fail early and predictably with bounded CLI errors, not rely on implicit falsy-string branch selection.

This is a narrow, low-risk maintainability hardening slice that complements the earlier source-validation proposals without duplicating their scope.

Smallest useful scope

• reject empty or whitespace-only --input values in both scripts with a clear argparse-facing error
• reject empty or whitespace-only --repo values in both scripts before any gh invocation
• choose fixture/live mode using explicit presence checks (is not None) after validation instead of string truthiness
• preserve existing valid fixture mode, valid live mode, and valid --fix behavior

Acceptance criteria

• scripts/check_bounty_issue_states.py rejects --input "", --input " ", and --repo " " with clear CLI validation and no Python traceback
• scripts/check_bounty_issue_states.py --fix continues to require a valid live --repo source and does not accept malformed empty source values
• scripts/check_live_bounty_closing_refs.py rejects --input "", --input " ", and --repo " " with clear CLI validation and no Python traceback
• valid fixture mode still works for both scripts
• valid live --repo owner/name mode still works for both scripts
• focused tests cover invalid empty/whitespace source values and at least one valid branch per script

Evidence

Current main source still shows the unfixed truthiness pattern:

• scripts/check_bounty_issue_states.py

  • required source group with source.add_argument("--input", ...) and source.add_argument("--repo", ...)
  • runtime selection: _load_input(args.input) if args.input else load_live_data(args.repo, args.api_host)
  • --fix gate still checks if args.input

• scripts/check_live_bounty_closing_refs.py

  • required source group with source.add_argument("--input", ...) and source.add_argument("--repo", ...)
  • runtime selection: _load_input(args.input) if args.input else load_live_data(args.repo, args.api_host, args.state, args.pr)

These are the same narrow failure shape already recognized in adjacent scripts, but for a different pair of bounty-check helpers.

Duplicate search

Checked existing open proposed-work issues and related submissions around:

• check_bounty_issue_states
• check_live_bounty_closing_refs
• source validation
• empty --input
• whitespace --repo

Related but distinct:

• #801 covers scripts/review_bounty_candidates.py
• #805 covers scripts/submission_quality_gate.py
• #807 covers the remaining report-script slice: scripts/proposed_work_triage.py, scripts/claim_inventory.py, and scripts/pr_queue_health.py
• #1104 covers wiring the closing-reference checker into PR readiness/CI, not its CLI source validation

I did not find an existing proposed-work issue specifically for source-argument hardening in check_bounty_issue_states.py and check_live_bounty_closing_refs.py.

Out of scope

• no change to bounty classification logic, closing-reference parsing, reopen policy, or live API semantics
• no payment execution, treasury mutation, wallet custody, bridge/exchange/off-ramp, or MRWK price behavior
• no claim that this proposed-work issue is itself an implementation bounty