docs(openspec): propose feedback-channels

[theCodeDrift]

Stack (root → tip):

• ➡️ docs(openspec): propose feedback-channels #477 (you are here)
  • feat(feedback): key surveys by kind and discriminate the payload on it #478
    • feat(feedback): send each kind to its survey, fill a bug's version info, and point to issues when telemetry is off #479
      • feat(agent): add the feedback and bug-report recipes, and rename the invited survey's recipe to rule-feedback #480
        • feat(skill): route feedback and bug reports to their recipes #481
          • docs(openspec): archive feedback-channels #482

Proposes splitting the agent feedback path into three channels, each sending to its own PostHog survey:

• agent rule-feedback: the invited rule-authoring survey (01a0c7b9…). It is renamed from agent feedback, and its questions, invite, cadence and dismiss are unchanged.
• agent feedback: general feedback (01a11da4…), listed in the agent index.
• agent bug-report: bug reports (01a11da7…), listed in the agent index. No GitHub account is needed.

feedback send --from takes a payload discriminated by a required kind (rule | general | bug) and validates it with zod. With telemetry off, nothing is sent and the CLI points the user to https://github.com/taskless/cli/issues.

This PR carries the stack's single changeset (patch, since the package is pre-1.0).

This stack lands with gh stack merge, which merges every PR to main in one all-or-nothing operation, so no layer reaches main on its own. That matters because the layers only work together. For example, #478 requires kind on the payload while the recipe the invite names still describes a single survey, and #480 lists bug-report in the index before the skill routes to it.

[theCodeDrift]

@claude /review

[github-actions]

Claude finished @theCodeDrift's task in 3m 28s —— View job

---

Review complete

Review mode: incremental — read 0 prior review thread(s) before reviewing. .prior-review.json had no reviewThreads, reviews, or comments beyond the trigger comment itself, so this is the first review of this PR; nothing to carry forward as addressed or still-open.

• Read .prior-review.json
• Fetch PR diff and metadata
• Read changed files in full (proposal, design, spec delta, tasks, changeset), plus the standing openspec/specs/cli-feedback-survey/spec.md and the current survey/constants.ts / schemas/feedback.ts / agent/ sources this change targets
• Review for correctness, consistency, and spec-delta integrity
• Post findings

This is a docs/OpenSpec-only PR (no package code touched), so no typecheck/lint/test run was needed or attempted.

What I checked specifically, given CLAUDE.md's warning that a ## MODIFIED Requirements block replaces the standing requirement wholesale and that a rename must be REMOVED + ADDED, never MODIFIED:

• Walked every MODIFIED Requirements block in specs/cli-feedback-survey/spec.md against the standing spec, scenario-by-scenario. All previously-existing scenarios are preserved verbatim or extended (cadence, feedback subcommand, payload-keys requirements) — nothing silently dropped.
• The rename of the invited recipe (feedback → rule-feedback) is correctly done as REMOVED + ADDED with a Migration note, not as a MODIFIED rename — exactly the pitfall the root CLAUDE.md calls out, and it's handled right here.
• Cross-checked the three PostHog survey IDs and question mappings between design.md's table and the spec's scenarios/payload tables — consistent throughout.
• Confirmed the changeset bump (patch) is correct per the root CLAUDE.md's pre-1.0 rule (this is added surface on a 0.y.z package).
• Confirmed the changeset lives on the bottom-of-stack PR per the stacking convention.

[New] One finding, posted inline on tasks.md lines 3–5: tasks 1.1–1.3 ("survey registry and payload schema") are added already checked off ([x]), but this PR's own diff contains no corresponding code, and the current packages/cli/src/survey/constants.ts / schemas/feedback.ts on this branch still reflect the old single-survey shape. That work belongs to a later layer in the stack (per the proposal's own "Delivery shape" section). Since this is the PR that first introduces tasks.md, shipping section 1 pre-checked is misleading to a reviewer looking at this PR in isolation — worth leaving those unchecked here and letting the PR that does the work check them off.

Nothing else stood out — the proposal, design, and spec delta are unusually careful about exactly the OpenSpec archive/rename hazards this repo has been burned by before.

[theCodeDrift]

Re: @github-actions — "Claude finished @theCodeDrift's task in 3m 28s…"
#477 (comment)

The one finding (tasks 1.1–1.3 pre-checked) is fixed in cb65006; those boxes are now left for the schema layer to check off.

— AI Coding Agent