Skills ralph
📦
ralph
v3.0.0 Content revision r1 Critical ⚙️ External commands🌐 Network access📁 Filesystem access
Run Specification-First Development Loops
Vague software requests often lead to rework and drift. Ralph helps Claude, Codex, and Claude Code clarify requirements, create specs, and verify progress through bounded loops.
Supports: Claude Codex Code(CC)
Do not auto-install this skill.
The canonical policy requires operator review before any installation action.
Install with my Agent
Copy this request to your Agent. It includes the canonical Skill page and manifest.
Agent request
Review the Skillstore skill "ralph" from https://skillstore.io/skills/supercent-io-ralph.md and its manifest at https://skillstore.io/api/skills/supercent-io-ralph/manifest. Verify the artifact. Do not auto-install. Inspect the skill and report your findings, then wait for an operator or manual installation decision.Your Agent should still show its plan and request any confirmation required by the security policy.
Agent-readable resources
Use these links when an AI agent, crawler, or script needs clean context instead of reading the full page.
Test it
Using "ralph". Interview me about a task management CLI.
Expected outcome:
Ralph asks focused questions about users, storage, commands, priority rules, and success criteria before suggesting any implementation.
Using "ralph". Fix all failing tests with a five iteration limit.
Expected outcome:
Ralph reports each attempt, lists remaining failures, records verification scores, and stops when tests pass or the limit is reached.
Using "ralph". Check whether this implementation drifted from the seed.
Expected outcome:
Ralph compares the result with the original goals, constraints, and ontology, then highlights areas that need correction.
Security Audit
Criticalv4 • 7/7/2026 Open versioned report
Most external-command findings are false positives caused by Markdown backticks, fenced examples, or literal prompt text. The confirmed risks are persistent writes to hidden agent configuration, Codex developer instruction injection, and broad Gemini hook guidance that can change agent behavior beyond one task.
3
Files scanned
935
Lines analyzed
24
Review items
0
False positives ignored
Confirmed security concerns (2)
Critical
Prompt Injection Attempt Detected
The setup script appends a new Codex developer_instructions contract with phrases like "Ouroboros specification-first workflow is active" and "continue immediately". This persists agent behavior outside the skill session.
The file directly defines the persistent instruction text and writes it into ~/.codex/config.toml. This is a clear agent-instruction persistence mechanism.
High
Wildcard AfterAgent Hook With Reduced Confirmation
SKILL.md instructs Gemini users to add an AfterAgent hook with matcher * that runs a bash command from ~/.gemini, then suggests sandbox plus yes mode to reduce prompts.
The hook configuration and reduced-confirmation command are shown together in the documented install flow. This creates broad automatic command execution risk if followed.
Capability review items (24)
These are real local capabilities that may be expected for this skill, so they require review but are not counted as confirmed malicious behavior.
High
Hidden file in home directory
# 1) developer_instructions → ~/.codex/config.toml (ooo command contract)
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file in home directory
# 2) ~/.codex/prompts/ralph.md (load via /prompts:ralph)
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file in home directory
# 3) ~/.codex/prompts/ouroboros.md (load via /prompts:ouroboros)
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file in home directory
echo " 1. Adds ooo command contract to ~/.codex/config.toml developer_instructions"
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file in home directory
echo " 2. Creates ~/.codex/prompts/ralph.md for /prompts:ralph"
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file in home directory
echo " 3. Creates ~/.codex/prompts/ouroboros.md for /prompts:ouroboros"
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file access
# 1) developer_instructions → ~/.codex/config.toml (ooo command contract)
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file access
# 2) ~/.codex/prompts/ralph.md (load via /prompts:ralph)
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file access
# 3) ~/.codex/prompts/ouroboros.md (load via /prompts:ouroboros)
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file access
echo " 1. Adds ooo command contract to ~/.codex/config.toml developer_instructions"
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file access
echo " 2. Creates ~/.codex/prompts/ralph.md for /prompts:ralph"
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file access
echo " 3. Creates ~/.codex/prompts/ouroboros.md for /prompts:ouroboros"
The line is descriptive text, but it names the hidden Codex home configuration targets that this setup script later creates or writes. Persisting agent prompts and developer instructions in ~/.codex is a real configuration risk.
High
Hidden file access
CODEX_DIR="$HOME/.codex"
This defines ~/.codex as the target directory and the script later writes config and prompt files under it. Hidden home-directory agent configuration can persist behavior across sessions.
High
Python file write/append
with open(path, "w") as out:
The Python write appends or rewrites Codex developer_instructions in ~/.codex/config.toml. Persistently changing agent instructions is a high-impact configuration risk.
High
Python file write/append
with open(path, "a") as out:
The Python write appends or rewrites Codex developer_instructions in ~/.codex/config.toml. Persistently changing agent instructions is a high-impact configuration risk.
High
Python file write/append
with open(path, 'a') as f:
The Python write appends or rewrites Codex developer_instructions in ~/.codex/config.toml. Persistently changing agent instructions is a high-impact configuration risk.
High
Hidden file in home directory
Required in `~/.gemini/settings.json`:
The documentation instructs users to modify hidden Gemini configuration or run hooks from ~/.gemini. This is user-directed, but it expands persistent agent hook and shell execution surface.
High
Hidden file in home directory
"includeDirectories": ["~/.gemini/extensions/ralph"]
The documentation instructs users to modify hidden Gemini configuration or run hooks from ~/.gemini. This is user-directed, but it expands persistent agent hook and shell execution surface.
High
Hidden file in home directory
AfterAgent hook for loop continuation (add to `~/.gemini/settings.json`):
The documentation instructs users to modify hidden Gemini configuration or run hooks from ~/.gemini. This is user-directed, but it expands persistent agent hook and shell execution surface.
High
Hidden file in home directory
"command": "bash ~/.gemini/hooks/ralph-check.sh",
The documentation instructs users to modify hidden Gemini configuration or run hooks from ~/.gemini. This is user-directed, but it expands persistent agent hook and shell execution surface.
High
Hidden file access
Required in `~/.gemini/settings.json`:
The documentation instructs users to modify hidden Gemini configuration or run hooks from ~/.gemini. This is user-directed, but it expands persistent agent hook and shell execution surface.
High
Hidden file access
"includeDirectories": ["~/.gemini/extensions/ralph"]
The documentation instructs users to modify hidden Gemini configuration or run hooks from ~/.gemini. This is user-directed, but it expands persistent agent hook and shell execution surface.
High
Hidden file access
AfterAgent hook for loop continuation (add to `~/.gemini/settings.json`):
The documentation instructs users to modify hidden Gemini configuration or run hooks from ~/.gemini. This is user-directed, but it expands persistent agent hook and shell execution surface.
High
Hidden file access
"command": "bash ~/.gemini/hooks/ralph-check.sh",
The documentation instructs users to modify hidden Gemini configuration or run hooks from ~/.gemini. This is user-directed, but it expands persistent agent hook and shell execution surface.
Risk Factors
⚙️ External commands (155)
scripts/setup-codex-hook.sh:79 scripts/setup-codex-hook.sh:163 scripts/setup-codex-hook.sh:164 scripts/setup-codex-hook.sh:166 scripts/setup-codex-hook.sh:170 scripts/setup-codex-hook.sh:171-182 scripts/setup-codex-hook.sh:182-186 scripts/setup-codex-hook.sh:186-195 scripts/setup-codex-hook.sh:195-198 scripts/setup-codex-hook.sh:198-206 scripts/setup-codex-hook.sh:206-212 scripts/setup-codex-hook.sh:212 scripts/setup-codex-hook.sh:213-214 scripts/setup-codex-hook.sh:214-215 scripts/setup-codex-hook.sh:215-216 scripts/setup-codex-hook.sh:216-221 scripts/setup-codex-hook.sh:221-229 scripts/setup-codex-hook.sh:229-258 scripts/setup-codex-hook.sh:258-262 scripts/setup-codex-hook.sh:262-268 scripts/setup-codex-hook.sh:268-269 scripts/setup-codex-hook.sh:269-270 scripts/setup-codex-hook.sh:270-271 scripts/setup-codex-hook.sh:271-272 scripts/setup-codex-hook.sh:272-273 scripts/setup-codex-hook.sh:273-274 scripts/setup-codex-hook.sh:274-275 scripts/setup-codex-hook.sh:275-279 scripts/setup-codex-hook.sh:279-284 scripts/setup-codex-hook.sh:284-288 scripts/setup-codex-hook.sh:288-291 scripts/setup-codex-hook.sh:291-309 scripts/setup-codex-hook.sh:309-310 scripts/setup-codex-hook.sh:310-311 scripts/setup-codex-hook.sh:311-312 scripts/setup-codex-hook.sh:312-313 scripts/setup-codex-hook.sh:313-317 scripts/setup-codex-hook.sh:317 scripts/setup-codex-hook.sh:318-319 SKILL.md:36-40 SKILL.md:40-46 SKILL.md:46-56 SKILL.md:56-67 SKILL.md:67 SKILL.md:68 SKILL.md:69 SKILL.md:70 SKILL.md:71 SKILL.md:72 SKILL.md:73 SKILL.md:74 SKILL.md:75 SKILL.md:76 SKILL.md:86-92 SKILL.md:92-96 SKILL.md:96-98 SKILL.md:98-103 SKILL.md:103-110 SKILL.md:110-113 SKILL.md:113-120 SKILL.md:120-124 SKILL.md:124-126 SKILL.md:126-129 SKILL.md:129-146 SKILL.md:146-150 SKILL.md:150-153 SKILL.md:153-157 SKILL.md:157-159 SKILL.md:159-168 SKILL.md:168-169 SKILL.md:169-170 SKILL.md:170-176 SKILL.md:176-179 SKILL.md:179-186 SKILL.md:186-206 SKILL.md:206-210 SKILL.md:210-211 SKILL.md:211-222 SKILL.md:222-226 SKILL.md:226-241 SKILL.md:241-245 SKILL.md:245-273 SKILL.md:273-279 SKILL.md:279-280 SKILL.md:280-281 SKILL.md:281 SKILL.md:287-290 SKILL.md:290-294 SKILL.md:294-299 SKILL.md:299-303 SKILL.md:303-310 SKILL.md:310-323 SKILL.md:323-329 SKILL.md:329-333 SKILL.md:333-336 SKILL.md:336-362 SKILL.md:362-369 SKILL.md:369-372 SKILL.md:372-373 SKILL.md:373-374 SKILL.md:374-375 SKILL.md:375-376 SKILL.md:376-384 SKILL.md:384-398 SKILL.md:398-400 SKILL.md:400-401 SKILL.md:401-402 SKILL.md:402-403 SKILL.md:403-405 SKILL.md:405-406 SKILL.md:406-414 SKILL.md:414-420 SKILL.md:420-431 SKILL.md:431-434 SKILL.md:434-435 SKILL.md:435-440 SKILL.md:440-442 SKILL.md:442-445 SKILL.md:445-446 SKILL.md:446-447 SKILL.md:447-448 SKILL.md:448 SKILL.md:452-457 SKILL.md:457-459 SKILL.md:459-460 SKILL.md:460-467 SKILL.md:467-469 SKILL.md:469-470 SKILL.md:470-483 SKILL.md:483-486 SKILL.md:486-488 SKILL.md:488-490 SKILL.md:490-491 SKILL.md:491-494 SKILL.md:494-496 SKILL.md:496 SKILL.md:497-505 SKILL.md:505-506 SKILL.md:506-507 SKILL.md:507-508 SKILL.md:508-516 SKILL.md:516-517 SKILL.md:517-518 SKILL.md:518-519 SKILL.md:519-520 SKILL.md:520-521 SKILL.md:521-522 SKILL.md:522-523 SKILL.md:523-524 SKILL.md:524-525 SKILL.md:525-526 SKILL.md:526-527 SKILL.md:527-528 SKILL.md:528-534 SKILL.md:534-548
🌐 Network access (6)
📁 Filesystem access (28)
scripts/setup-codex-hook.sh:5 scripts/setup-codex-hook.sh:6 scripts/setup-codex-hook.sh:7 scripts/setup-codex-hook.sh:29 scripts/setup-codex-hook.sh:30 scripts/setup-codex-hook.sh:31 scripts/setup-codex-hook.sh:5 scripts/setup-codex-hook.sh:6 scripts/setup-codex-hook.sh:7 scripts/setup-codex-hook.sh:29 scripts/setup-codex-hook.sh:30 scripts/setup-codex-hook.sh:31 scripts/setup-codex-hook.sh:42 scripts/setup-codex-hook.sh:55 scripts/setup-codex-hook.sh:90 scripts/setup-codex-hook.sh:98 scripts/setup-codex-hook.sh:144 scripts/setup-codex-hook.sh:119 scripts/setup-codex-hook.sh:125 scripts/setup-codex-hook.sh:141 SKILL.md:459 SKILL.md:464 SKILL.md:469 SKILL.md:477 SKILL.md:459 SKILL.md:464 SKILL.md:469 SKILL.md:477
Audited by: codex View Audit History →
Share & cite this report
Share the versioned assessment report, neutral badge, embed card, and citations. Skillstore reports evidence without deciding whether this Skill is safe.
Copy report link
https://skillstore.io/skills/supercent-io-ralph/audits/4?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown badge
[](https://skillstore.io/skills/supercent-io-ralph?utm_source=security_passport_badge)HTML badge
<a href="https://skillstore.io/skills/supercent-io-ralph?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/supercent-io-ralph/security.svg" alt="Skillstore security assessment" loading="lazy"></a>Embed card
<iframe src="https://skillstore.io/embed/skills/supercent-io-ralph.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>Academic citations (APA · BibTeX · CFF)
APA citation
supercent-io. (2026). ralph security audit report (audit version 4) [Author version 3.0.0]. Skillstore. https://skillstore.io/skills/supercent-io-ralph/audits/4BibTeX citation
@techreport{supercent-io-supercent-io-ralph-2026,
author = {supercent-io},
title = {ralph security audit report (audit version 4)},
institution = {Skillstore},
year = {2026},
number = {4},
url = {https://skillstore.io/skills/supercent-io-ralph/audits/4},
note = {Author version 3.0.0}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "ralph security audit report (audit version 4)"
version: "3.0.0"
type: report
authors:
- name: "supercent-io"
date-released: "2026-07-07"
url: "https://skillstore.io/skills/supercent-io-ralph/audits/4"
identifiers:
- type: other
value: "skillstore:supercent-io-ralph:audit:4"
description: "Skillstore immutable audit report identifier"
Skillstore Score
Why this score Evidence Confidence: Medium 45
Architecture
100
Maintainability
87
Content
69
Community
87
Spec Compliance
What You Can Build
Clarify a Product Request
Use a Socratic interview to turn a vague feature idea into explicit goals, constraints, and acceptance criteria.
Fix a Failing Build
Run a bounded Ralph loop that attempts fixes, verifies results, records failures, and stops at success or the iteration cap.
Control Project Drift
Compare implementation progress against the original seed specification and decide when correction is needed.
Try These Prompts
Start an Interview
Use Ralph to interview me about this idea before writing code: [describe the product or feature].
Create a Seed Spec
Create a Ralph seed specification from our interview. Include goals, constraints, acceptance criteria, and open questions.
Run a Bounded Fix Loop
Run Ralph on this task with a maximum of five iterations: fix the failing checks and report verification after each iteration.
Evaluate and Evolve
Evaluate the current implementation against the seed spec, identify drift, and propose the next evolution step with a stop condition.
Best Practices
- Set an explicit maximum iteration count before starting any Ralph loop.
- Review generated setup scripts before allowing changes to agent configuration files.
- Keep the seed specification short, testable, and tied to observable acceptance criteria.
Avoid
- Do not use Ralph to bypass user review for risky file, shell, or network actions.
- Do not start long loops from unclear goals or missing acceptance criteria.
- Do not install global hooks without understanding the commands they run.
Frequently Asked Questions
What does Ralph do?
Ralph structures coding work through interviews, seed specifications, execution, verification, and controlled iteration.
Does Ralph write code by itself?
It guides the agent through coding tasks, but user permissions and tool policies still control actual file and command actions.
Why does it ask questions first?
The interview reduces ambiguity before implementation, which can prevent rework and missed requirements.
Can I use it with Codex?
Yes. The skill includes Codex guidance, but its setup script can modify local Codex configuration.
Is the loop unlimited?
No. The skill documents default and custom iteration limits, and users should keep those limits explicit.
What should I review before installing?
Review any script that writes to hidden agent config files or registers hooks that run shell commands.
Developer Details
Author
supercent-ioLicense
MIT
Author version
v3.0.0
Skillstore revision
r1
Ref
f93e9bb0daca99badb6a7e574b97737155d57cb3
Maintenance freshness
7/18/2026
Usage
22 downloads · 146 views
File structure