Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
npx agentmods add commands/transilienceai/communitytools/pendinggit clone --depth 1 https://github.com/transilienceai/communitytoolsWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/commands/transilienceai/communitytools/pending)<a href="https://agentmods.dev/commands/transilienceai/communitytools/pending"><img src="https://agentmods.dev/badge/commands/transilienceai/communitytools/pending.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5 | $0.00043 | $0.00577 |
| Opus 5 | $0.00022 | $0.00289 |
| Sonnet 5 | $0.00009 | $0.00115 |
| Haiku 4.5 | $0.00004 | $0.00058 |
Grade A, and why
pending scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 5d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 54 lines — stays where its author put it; the contents beside it link to each section on GitHub.
/pending — Curator commands
Arguments: <action> [<candidate_id>] where action is one of list, review, promote, reject, hold.
What to do
/pending list
- Read
.claude/pending/*.yaml. - Print a table:
Learning candidates (pending curator review):
ID Status Target detector Proposed change
pending-2026-04-19-nonprod-severity-downgrade proposed any-any-broadness Downgrade severity for non-prod assets
pending-2026-04-20-ssh-management-whitelist proposed public-source-allow Allow public-source if documented bastion
...
/pending review <candidate_id>
- Read
.claude/pending/<candidate_id>.yaml. - Pretty-print the full candidate: proposal_type, target_skill, change_summary, evidence (finding_ids + auditor_reasons), test_cases, validation_status.
- At the end, print:
Decide with: /pending promote <id> | /pending reject <id> | /pending hold <id>
/pending promote <candidate_id>
- Set
validation_status: approvedin the candidate YAML. - Create a new git branch:
learning/<candidate_id>. - Tell the operator: "Branch
learning/<candidate_id>created. Open../detectors/<target_skill>.mdand apply the proposed change, add regression golden tests, commit, and open a PR." - Do NOT actually edit the skill file yourself — promotion is a human curator action; you only prepare the branch and point the way.
/pending reject <candidate_id>
- Set
validation_status: rejectedin the candidate YAML. - Ask the operator for a reason and record it in
curator_notes.
/pending hold <candidate_id>
- Set
validation_status: heldin the candidate YAML. - Ask the operator for a reason (e.g., "need more engagement data") and record it in
curator_notes.
Hard rules
- Never auto-promote. Curator always reviews. This is the "do not learn blindly" guarantee.
- Preserve the audit trail —
curator_notesshould always be filled on reject/hold. - Promotion creates a branch but does NOT modify the skill. The human reviews the proposed change, writes regression tests, and lands the PR manually. This ensures every skill update has evidence + tests before it ships to 1001+ auditors.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 5d ago First seen · 54 lines · 43 tokens per session scan A 16a19bb1cb7c
pending is a command published in the GitHub repository transilienceai/communitytools (507 stars, last pushed 1mo ago), licensed MIT. It adds 43 tokens to every session and 577 once invoked, about $0.0002 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other commands, from other repositories
brain
Manage the engagement brain. Subcommands: 'init' to set up, 'brief ' for pre-flight, 'status' for overview, 'exhausted [target]' to see dead ends.
chain
Build deep exploit chains — dispatches chain-builder agent. Given bug A, recursively walks the chain graph. Usage: /chain (then describe bug A).
correlate
Run the finding correlation engine to discover attack chains from individual findings.
fullscan
Full security assessment with brain coordination. Multi-phase, skips known-exhausted areas, builds on prior knowledge.
learn
Record a platform response and update learning. Usage: /learn [--bounty 500] [--vuln-type XSS].
mindmap
Generate a text-based attack surface mindmap. Shows tech stack → vuln class → endpoint relationships. Usage: /mindmap.