agentsclimarketplace

Rejection recovery loop

Skill mturac/everything-openai-codex/skills/rejection-recovery-loop

Turn public launch, directory, community, or list rejections into repo fixes and better proof without arguing, spamming, or resubmitting blindly. Use after Hacker News, Product Hunt, GitHub list PR, marketplace, or community rejection.From its SKILL.md

Install
npx -y skills add mturac/everything-openai-codex --skill rejection-recovery-loop

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

SKILL.md

4.8 KB, 983 tokens by cl100k_base, as published. Nobody here has run it

Rejection Recovery Loop

Use this skill when a public submission, listing PR, community post, launch surface, or marketplace attempt is rejected, closed, ignored, or blocked.

The goal is not to "win the argument." The goal is to learn exactly what failed, fix the project surface when the criticism is valid, and decide whether to resubmit, wait, or stop.

Rule Precedence

Follow this order when instructions conflict:

  1. Safety, privacy, credentials, and target-platform rules
  2. No-bypass and no-spam boundaries in this skill
  3. Exact rejection evidence
  4. User-requested distribution goals
  5. Copy style and timing preferences

Public-action carve-outs: zero inside this skill for reposting, new-account workarounds, duplicate PRs, public replies, form submissions, or account changes. This skill always stops at a recovery report and optional response draft.

Allowed action carve-outs: zero inside this skill.

Conflict behavior: when a requested action conflicts with the ordered rules above, stop after the recovery report and name the conflicting rule.

This skill does not define or call tools directly. If tools are needed, use only the active runtime tool schema and treat missing authentication, hidden moderation state, or unclear rejection text as residual risk.

Blocked Actions

  • arguing with maintainers or moderators
  • reposting the same submission with cosmetic wording changes
  • creating a new account, new PR, or new thread to bypass a rejection
  • treating silence as permission to ping repeatedly
  • hiding lineage, upstream adaptation, or project maturity

Technical Risk Contract

Before editing repo files, identify the rejected surface, target repo files, and rollback path.

Allowed edits from this skill are limited to durable recovery surfaces such as README copy, attribution notes, screenshots, public-review docs, release notes, skills, and tests that verify the corrected claim.

Excluded technical changes from this skill:

  • product behavior
  • form submissions
  • posted comments
  • new listing PRs
  • account changes

Verification must include the relevant docs tests, catalog checks when counts change, PromptGuard for prompt-like files, and a final diff summary.

Classification

Classify the rejection before taking action:

TypeSignalMeaning
Account gatepermission or eligibility blockaccount-state blocker
Category mismatchwrong list, wrong section, not enough fittarget-fit blocker
Proof gapasks for users, screenshots, validation, install proofevidence blocker
Quality gapbroken README, unclear install, bad screenshots, CI failurerepo-quality blocker
Duplicateexisting PR, existing item, replacement threadduplicate blocker
Policy gatecommercial, affiliate, AI-spam, self-promo rulespolicy blocker
Maintainer requestconcrete requested changeresponse-worthy blocker
Silent closeclosed without reasonunverifiable close

Workflow

  1. Capture evidence.
    • Save the exact rejection text, URL, timestamp, and actor if available.
    • Link the submitted copy, PR body, screenshots, and target rules.
  2. Find the smallest truthful fix.
    • README clarity
    • install command accuracy
    • screenshots or demo evidence
    • category rationale
    • attribution and upstream lineage
    • CI, audit, or package evidence
  3. Decide the next move.
    • fix and respond: maintainer asked for a concrete change.
    • fix and wait: repo surface was weak but no response is needed.
    • resubmit later: target fit is real, but proof is not ready.
    • skip permanently: target rules or audience fail to fit.
  4. Produce a response draft only when response is justified.

Response Rules

A response is justified only when:

  • the maintainer asked a concrete question
  • a broken link or factual issue was fixed
  • a CLA or category instruction needs one clear answer
  • the target explicitly invites resubmission after changes

Keep it to one comment. Cite the fix. Avoid asking for special treatment.

Output

Return exactly these fields. Status must be one of fix and respond, fix and wait, resubmit later, or skip permanently.

## Rejection Recovery

Status: fix and respond | fix and wait | resubmit later | skip permanently

### Evidence
| Item | Link or Quote | Meaning |
| --- | --- | --- |

### Classification
[type and reason]

### Repo Fixes
- [specific file, asset, CI, or docs fix]

### Response Draft
[only if justified]

### Residual Risk
- [moderator discretion, maturity gap, missing adoption proof]

What ships with it

Read from the repository

Just SKILL.md. No reference files, no scripts.

Keep looking

Skills are one crate of 326,852. Ordering is by how many stacks a row turns up in, so the top of any crate is what has actually been picked rather than what has the most stars.