docs(parallel-issues): warn that gh api -f/-F promotes reads to POST

[thewrz]

This was written agentically; verify its assertions:

Why

gh api infers its HTTP method from its flags: any -f/-F parameter promotes the request to POST unless -X GET is explicit. The kit's REST-first guidance told agents to route bulk reads over gh api repos/... without saying so, and a field run's filtered issue read became an attempted issue create in a repository being read. The knowledge lived only in helper code (merge-gate.sh passes -X GET); it now lives at the instruction sites agents actually read.

What

• parallel-issues/references/triage-and-selection.md: a short paragraph after the REST-routing rule stating the -f/-F → POST promotion, plus a correct filtered-read example (gh api -X GET "repos/$REPO/issues" -f labels=bug -f state=open).
• parallel-issues/SKILL.md: one clause on its own REST-routing sentence and a pointer to the example — kept minimal because the file sits at its size ceiling.
• No guard, hook, or lint added (issue's explicit out-of-scope list); no ceiling raise needed.

Decisions

• agentkit/references.md from the issue's scope list does not exist; the real index (agentkit/skills/references.md) only points at triage-and-selection.md, so it needed no edit.
• Example uses $REPO rather than a literal <owner>/<repo> placeholder: shellcheck (via the markdown-block gate) misparses the angle brackets as redirection.
• Root review: diff within the pinned write set; correctness/security/tests lenses clear.

Diff facts (root, diff-facts.sh --base origin/main):
base=origin/main
files=2
total.insertions=14
total.deletions=3
total.lines=17
operational.files=2
operational.insertions=14
operational.deletions=3
operational.lines=17
generated.files=0
generated.insertions=0
generated.deletions=0
generated.lines=0
lockfile.files=0
lockfile.insertions=0
lockfile.deletions=0
lockfile.lines=0
fixture.files=0
fixture.insertions=0
fixture.deletions=0
fixture.lines=0
non_operational.files=0
non_operational.insertions=0
non_operational.deletions=0
non_operational.lines=0

Testing

• Worker: agent-run.sh --cmd test ALL GREEN (.agent/logs/20260825T133432Z-test.log)
• CI green
• Adversarial review (Codex gpt-5.6-sol, xhigh) — receipt posted
• Manual: grep both instruction sites for -X GET

🤖 Co-authored by Claude Sonnet 5 (worker) / Claude Fable 5 (root).

Closes #443

[coderabbitai]

Important

Review skipped

Auto reviews are disabled on this repository. Please check the settings in the CodeRabbit UI or the .coderabbit.yaml file in this repository. To trigger a single review, invoke the @coderabbitai review command.

⚙️ Run configuration

Configuration used: Organization UI

Review profile: CHILL

Plan: Pro Plus

Run ID: 207e0324-9804-47fa-8f37-b333dd4b19c8

You can disable this status message by setting the reviews.review_status to false in the CodeRabbit configuration file.

Use the checkbox below for a quick retry:

• 🔍 Trigger review

---

_{Comment @coderabbitai help to get the list of available commands.}

[thewrz]

This was written agentically; verify its assertions:

Adversarial review receipt

• Reviewer: provider=openai; model=gpt-5.6-terra; effort=xhigh; mode=cross-provider (reason: peer-cli=codex present; runner-pinned model (operator requested gpt-5.6-sol; no override exists, see The adversarial reviewer cannot be declared, though the worker's model, effort, and fallback can #452))
• Counts: P1=0; P2=0; total=0
• Confirmed finding: none confirmed

🤖 Co-authored by Claude Fable 5 (root orchestrator).