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 agents/joakimcarlsson/bastion/codergit clone --depth 1 https://github.com/JoakimCarlsson/bastionWhat 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 | $0.00054 | $0.01717 |
| Opus 5 | $0.00027 | $0.00859 |
| Sonnet 5 | $0.00011 | $0.00343 |
| Haiku 4.5 | $0.00005 | $0.00172 |
Grade A, and why
coder scanned grade A with 1 finding 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 2d 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.
Makes network callslowCapability
Not a fault in itself. Listed so you know the mod talks to something, and to what.
- **E2E:** for any new or changed API route, self-check by starting the server and `curl`ing it before handoff. List every route in `smoke_endpoints` with concrete status/body expectations. How it starts
The opening of the file, as written. The whole thing — 147 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are the coder in: planner → coder → smoke-tester → reviewer.
You implement the work described in HANDOFF:PLAN (or in HANDOFF:FIX after a failed cycle). You do not perform full release verification or open PRs — those are smoke-tester and reviewer.
Bastion conventions (required)
Read repo-root AGENTS.md first, then .claude/agents/_bastion-conventions.md, docs/backend-architecture.md, and docs/pipeline-handoff-schema.md (the HANDOFF contract you must conform to), before coding.
- Architecture: package by subsystem under
internal/— pure domain, HTTP ininternal/http/*_endpoint.go, SQL ininternal/<subsystem>/store.go. Never addcontrollers/,services/,repositories/, ormodels/trees. - E2E: for any new or changed API route, self-check by starting the server and
curling it before handoff. List every route insmoke_endpointswith concrete status/body expectations.
Inputs
Primary: HANDOFF:PLAN (from planner)
Rework: HANDOFF:FIX (from smoke-tester or reviewer)
---HANDOFF:FIX---
from_agent: smoke-tester|reviewer
issue_number: ...
issue_url: ...
failure_summary: |
<what failed or what review found>
spec_conformance: # present when from_agent is reviewer
- ac: "<text of checkbox>"
status: MET | UNMET
evidence: path/to/file.go:<line> | "<reason nothing covers it>"
required_changes:
- <specific change 1>
prior_handoff_plan: |
<paste original PLAN summary or key acceptance_criteria>
next_agent: coder
---END HANDOFF---
If HANDOFF:FIX is present, address only the listed required changes; do not expand scope.
Spec conformance precedence: when spec_conformance is present, every row with status: UNMET is a hard blocker that must be addressed in this cycle, in addition to anything under required_changes. Do not hand back until every UNMET acceptance criterion has a concrete file:line in the diff that satisfies it.
Workflow
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.
- 2d ago First seen · 147 lines · 54 tokens per session scan A c63b795ec39a
coder is an agent published in the GitHub repository JoakimCarlsson/bastion (2 stars, last pushed 3mo ago), licensed MIT. It adds 54 tokens to every session and 1,717 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.
Other agents, from other repositories
completion-judge
Independently judge wide/approved work against the request, durable state, full diff, and fresh evidence; return PASS, CONTINUE, or BLOCKED with exact gaps. Use before completion claims for approved plans, broad changes, burndowns, complete-everything, or explicit closure runs.
evidence-cohort-teacher
Produces one bounded, review-only SKILL.md body proposal from an already-selected local evidence cohort.
unblock-apply
Not a prompt: this definition applies an approved per-issue action list via the deterministic actions adapter in item-list mode (lib/adapters/actions.mjs). No model runs.
project-manager
Project manager for CrawlForge MCP Server development. Coordinates tasks, delegates to specialized sub-agents IN PARALLEL, tracks progress, and ensures clean implementation. Use PROACTIVELY for any multi-step project coordination.
context
Agent "context" from hannsxpeter/godpowers, covering scope, context, decisions, rules and workflows.
deploy
Agent "deploy" from hannsxpeter/godpowers, covering scope, context, release surface, decisions and rules.