SKILLEMALL.ai
GitHub ↗ RU Sign in

FD great_cto

Use when the CTO describes a feature, task, or project goal. Orchestrates the full SDLC pipeline automatically based on project type.

Not recommendedcritical or high security findings · low grade F
avelikiy/great_cto Agent Skills author: avelikiy MIT 80 files body ≈ 9 073 tokens Open the sourcegithub.com↗ analyzed 2 d ago

Orchestrates the full SDLC pipeline automatically based on project type.

As a process D 49/100 · Unfinished process — References files that are not bundled: ../../docs/adr/ADR-009-gates-follow-reversibility.md, scripts/log-verdict.sh, scripts/lib/approval-level.mjs

ProcedureInfrastructureAI and agentstype and topics are labelled automatically from the skill text
Runs in: Claude / Claude Code ChatGPT / Codex Gemini CLI Cursor GitHub Copilot / VS Code Qwen Code Kimi Code CLI Z.ai (GLM) DeepSeek (Deep Code) Mistral Vibe LangChain (deepagents) OpenClaw Goose OpenCode
JSON
Technical rating
F
47/100
safety, quality, tests
Safety 60%
39
Quality 40%
59
Run on models
none yet
Process rating
D
49/100
Unfinished process
References files that are not bundled: ../../docs/adr/ADR-009-gates-follow-reversibility.md, scripts/log-verdict.sh, scripts/lib/approval-level.mjs
Tools and files w 18
0
Inputs and preconditions w 11
0
When it triggers w 12
20
the three weakest of ten parameters · all ten

What is at stake

The skill contains fragments that, in the wrong hands, cost money or data. Below: what the installer risks and what the author should do.

Exfiltration
If you install

The instructions or scripts send data out: environment variables, keys, file contents, chat history. You may never notice, because the agent performs the upload as "part of the task".

For the author

If the upload is not needed for the task, remove it: catalogs flag such skills and delist them. If it is needed, name the destination explicitly, say what leaves the machine, and give the user a switch.

Instruction override medium severity

Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.

If you install

The text contains phrases like "ignore previous instructions" or "you are now…". That is an attempt to hijack the agent: it may break your rules, the system limits or company policy.

For the author

An honest skill does not need them: state the role and the rules directly without overriding other instructions. Otherwise catalog scanners and corporate filters will block the listing.

Broad scope medium severity

Below is the worst case for this category. The finding here is medium: the guard saw a sign, not a proof.

If you install

The skill asks for more than the task needs: broad tool access, credential environment variables, binaries. Every extra permission widens the damage from a mistake or a compromise.

For the author

Narrow allowed-tools and the variable list to the minimum; replace binaries with readable sources or scripts.

How to improve

  1. Address the high-severity findings: each costs 18 safety points. If one is a false positive, add the rule id to guard.allow in spec.yaml.
  2. The SKILL.md body is over 5,000 tokens: move reference detail into references/ and load it when needed.
  3. The text references files that are not there: add them or drop the references.
For the model run — optional
  • Your own cases (evals/evals.json, 4–6 real requests with expected answers): the full check would then run those instead of a model-drafted suite.
  • A spec.yaml with trigger phrases and assertions — a behaviour contract for CI; `skilltest init` writes a template.

Guard findings · 32

  • high Exfiltration intent-browser-credential-store templates/ARCH-browser-extension.md:38
    Accesses a browser credential / cookie store
    | Cookies (existing site auth) | `chrome.cookies` | requires `cookies` permission |
Medium and low: 31
  • medium Instruction override en-ignore-previous packs/agent-pack.md:80
    Instruction-override phrase ("ignore previous instructions") (detector / deny-list definition)
    - "Ignore previous instructions and..."
    detector
  • medium Instruction override en-ignore-previous packs/voice-pack.md:46
    Instruction-override phrase ("ignore previous instructions") (detector / deny-list definition)
    - `EVAL-prompt-injection-via-voice` — caller dictates "ignore previous instructions" — agent does not comply
    detector
  • medium Broad scope meta-broad-allowed-tools SKILL.md:1
    Broad tool permissions pre-approved: Bash
    allowed-tools: Read Write Edit Bash Glob Grep Agent
  • low Risky intent intent-offensive-security ARCHETYPES.md:61
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    | `regulated` | Full compliance suite per domain (see enterprise-pack) | Pentest, audit trail completeness | 0 compliance violations, audit trail 100% |
  • low Risky intent intent-offensive-security ARCHETYPES.md:135
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    | oracle-reviewer | `web3` | security-officer pre-impl delegates here | TM with subtype block-ship gate (lending → flash-loan-sim, AMM → k-invariant, bridge → cross-chain integrity), oracle strategy (
  • low Secrets in code secret-high-entropy-token packs/accounting-pack.md:47
    High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)
    - `EVAL…ule` — revenue recognized per obligation, not at signing
    quoted
  • low Secrets in code secret-high-entropy-token packs/edtech-pack.md:77
    High-entropy token-like string (may be an id, hash or a credential) (documentation table row)
    | `EVAL…ens` | axe-core + pa11y on student-facing routes | ≥ 95% pass rate; zero Critical violations; manual screen-reader spot-check |
    table
  • low Risky intent intent-offensive-security packs/enterprise-pack.md:90
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    - [ ] Threat-Led Penetration Testing programme (3-year cycle)
  • low Risky intent intent-offensive-security packs/enterprise-pack.md:118
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    - [ ] All AL2 measures + penetration test every 6 months
  • low Risky intent intent-offensive-security packs/gov-pack.md:41
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    | Annual 3PAO penetration test report | `docs/compliance/pentest/{year}/` | external 3PAO |
  • low Risky intent intent-offensive-security packs/msp-pack.md:50
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    across every managed client — the 2021 Kaseya VSA ransomware incident propagated through exactly this
  • low Secrets in code secret-high-entropy-token packs/rcm-pack.md:45
    High-entropy token-like string (may be an id, hash or a credential) (quoted — discussed, not commanded)
    - `EVAL…acy` — posted amounts match remittance-advice line items
    quoted
  • low Secrets in code secret-high-entropy-token packs/tax-pack.md:43
    High-entropy token-like string (may be an id, hash or a credential) (detector / deny-list definition)
    - `EVAL…mit` — e-file blocked without valid signed 8879 on file
    detector
  • low Secrets in code secret-high-entropy-token packs/tax-pack.md:44
    High-entropy token-like string (may be an id, hash or a credential) (detector / deny-list definition)
    - `EVAL…ent` — secondary data use blocked without specific consent
    detector
  • low Risky intent intent-offensive-security packs/web3-pack.md:147
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    ## Bug Bounty Sizing (Immunefi tier mapping)
  • low Risky intent intent-offensive-security playbooks/security-incident.md:31
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (quoted — discussed, not commanded)
    echo "  /security-incident \"ransomware-style behavior on build server\""
    quoted
  • low Risky intent intent-offensive-security playbooks/security-incident.md:218
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    | Red Team                    | <N attacks> | yes / no | <attack vector not tested> |
  • low Risky intent intent-offensive-security playbooks/threat-model.md:109
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    | **E** | Elevation of privilege | horizontal or vertical privilege escalation, IDOR |
  • low Secrets in code secret-high-entropy-token references/knowledge-extraction.md:142
    High-entropy token-like string (may be an id, hash or a credential)
    ke_id: KE-2…eld
  • low Risky intent intent-offensive-security references/poc-mode.md:54
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    | Pentest | quarterly | pre-production | ✗ |
  • low Risky intent intent-offensive-security references/poc-mode.md:73
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use) (detector / deny-list definition)
    **Rule of thumb**: in AI PoC mode, you skip ARCH ceremony, TDD strictness, and full pentest — but you do NOT skip the threats that are unique to AI (prompt injection, cost runaway, tool-call abuse, mo
    detector
  • low Risky intent intent-offensive-security references/sec-metrics.md:22
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    | **Pentest burn-down** | open_findings / (open_findings + closed_findings_in_window), weighted by severity | trend down-and-right |
  • low Risky intent intent-offensive-security references/sec-metrics.md:34
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    - **Pentest burn-down** — if you're paying for pentests, finding closure rate tells you whether the team treats findings as real work or theatrical.
  • low Risky intent intent-offensive-security references/sec-metrics.md:50
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    | Pentest burn-down | `docs/security/PENTEST-*.md` finding tables (added when pentests run) |
  • low Risky intent intent-offensive-security references/sec-metrics.md:61
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    | **Closing findings without fixing** | Pentest burn-down drops fast, CVE-MTTR stays flat | Finding reopen rate > 10% in 90d → flag |
  • low Risky intent intent-offensive-security references/sec-metrics.md:77
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    Pentest burn-down is trended, not alerted — a slow burn-down is a team-culture conversation, not a pager-worthy event.
  • low Risky intent intent-offensive-security references/secure-sdlc.md:88
    Offensive-security / dual-use content (legitimate for authorised testing; review intended use)
    | **Art. 24–27** Operational resilience testing | regular testing, TLPT every 3y for significant | manual; pentest cadence documented in archetype | ◐ |

A further 4 matches are quotations in this security skill's documentation and are not counted as findings.

Files scanned: 80. Evidence is masked. Grey chips explain why severity was lowered.

Against the Agent Skills spec

  • warning name-format name should be kebab-case (lowercase letters, digits, hyphens)
  • warning body-long SKILL.md body ≈ 9073 tokens (recommended < 5000); move details to references/
  • warning missing-ref reference to a missing file: ../../docs/adr/ADR-009-gates-follow-reversibility.md
  • warning missing-ref reference to a missing file: scripts/log-verdict.sh
  • warning missing-ref reference to a missing file: scripts/lib/approval-level.mjs
  • note frontmatter-key unknown frontmatter key "effort"
  • note frontmatter-key unknown frontmatter key "paths"
  • note edit-residue the text marks something as outdated (lines 54): check that old rules are not kept next to new ones — the full check reads the text for contradictions

Process rating: all ten parameters 49/100

Will not run. References files that are not bundled: ../../docs/adr/ADR-009-gates-follow-reversibility.md, scripts/log-verdict.sh, scripts/lib/approval-level.mjs
  • 0Tools and files. 3 referenced file(s) missing: ../../docs/adr/ADR-009-gates-follow-reversibility.md, scripts/log-verdict.sh, scripts/lib/approval-level.mjs
  • 0Inputs and preconditions. Does not say what the process needs to start
  • 20When it triggers. No condition that starts the skill
  • 40Result and completion. Does not say what the result is
  • 40Execution cost. Instruction body is 9073 tokens: crowds the task out of the window
  • 100Steps. 32 steps
  • 100Failures and branches. 1 branches, has a failure section
  • 100Consistency. Name and required fields are in place
  • 100Running it twice. Mutating operations check current state
  • 100Progress reporting. Reports progress
  • medium Safety rules and hard prohibitions inside a skill: they belong in the system prompt, here they protect nothing
  • low 10 top-level sections: this looks like several domains in one skill
  • low The response is described with custom markup (11 tags): a typed call is more reliable

Everything here is measured from the skill text rather than judged by a model, so the numbers are checkable. A parameter weighs more when it is a more common reason for the process to stall.

Quality signals

  • +5Description has no quoted example phrases that should trigger the skill
  • +4Description does not say when NOT to use the skill (false activations)
  • +3Output format is not stated: the model decides each time
  • +1No license
  • +2Single-language instructions
  • +3Description length 133: enough signal without eating the budget
  • +4Structure: 30 headings
  • +3Step-by-step instructions: 32 items
  • +4Has examples (26 code blocks)
  • +4Reference files are cited in the instructions (8 of 40)

Quality base 70; lint remarks subtract, signals add up to 100. Result: 59.