Add aeon skills#366
Merged
Merged
Conversation
These skills come from aeon (https://github.com/aaronjmars/aeon), the autonomous agent framework. Ports cover four areas that complement the existing Bankr catalog: Agent meta-tooling (5) - aeon-autoresearch — genetic-algorithm-style skill self-improvement - aeon-skill-repair — auto-diagnose and fix failing skills, systemic-first - aeon-skill-evals — assertion-based regression detection across runs - aeon-skill-security-scan — injection/exfiltration/traversal/prompt-override audit - aeon-spawn-instance — clone agents into new GitHub repos with tailored skill sets Multi-agent ops (2) - aeon-fleet-control — parent-side control plane: status, dispatch, aggregate - aeon-operator-scorecard — weekly "was this week worth it?" synthesis Security (2) - aeon-vuln-scanner — responsible-disclosure scanner with PVR pipeline - aeon-workflow-security-audit — zizmor + actionlint with delta classification On-chain payments (1) - aeon-distribute-tokens — Bankr Wallet API batch payouts with idempotency Research and signal (8) - aeon-narrative-tracker — quantitative mindshare + velocity + position calls - aeon-unlock-monitor — Absorption Ratio supply-pressure tracker - aeon-deep-research — multi-source synthesis with credibility tiering - aeon-defi-overview — daily regime read + sustainable-vs-incentive yields - aeon-monitor-runners — top runners across chains via GeckoTerminal - aeon-last30 — cross-platform 30-day social research - aeon-token-pick — disciplined token + prediction-market picks with skip branch - aeon-reg-monitor — legislation/regulation tracker, Stage × Impact triage - aeon-monitor-polymarket — watchlist-driven prediction-market monitor Marketing (1) - aeon-schedule-ads — multi-platform ad scheduler, PAUSED by default 🤖 Generated with Claude Code Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
After comparing against the existing Bankr catalog, the meta-tooling skills (autoresearch, spawn-instance, skill-repair, skill-security-scan, skill-evals, fleet-control, operator-scorecard, workflow-security-audit, schedule-ads) only function inside the aeon framework — they reference state files, issue trackers, and workflow patterns that don't exist for a general Bankr user. Removed them and added 9 standalone consumer skills: Removed (aeon-framework-coupled or niche) - aeon-autoresearch, aeon-spawn-instance, aeon-skill-repair - aeon-skill-security-scan, aeon-skill-evals - aeon-fleet-control, aeon-operator-scorecard - aeon-workflow-security-audit (CI-only, narrow) - aeon-schedule-ads (AdManage.ai dependency, niche) Added (work out-of-the-box, useful for crypto+AI consumers) - aeon-token-movers — CoinGecko movers/losers with pump-risk flags - aeon-on-chain-monitor — watchlist for addresses/contracts, RPC-driven - aeon-defi-monitor — pools/lending/vaults watchlist with health checks - aeon-monitor-kalshi — Kalshi watchlist + cross-venue arb vs Polymarket - aeon-paper-pick — one AI paper to read today from HF Papers - aeon-huggingface-trending — trending HF models/datasets/spaces, filtered - aeon-hacker-news-digest — top HN with comment-mined insights - aeon-rss-digest — multi-feed daily roll-up with cross-feed dedup - aeon-deal-flow — weekly funding round tracker, primary-source-required Also cleaned up `memory/` path references in aeon-unlock-monitor and aeon-vuln-scanner so they read as standalone state files rather than aeon- framework paths. 🤖 Generated with Claude Code Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
The earlier removal of these skills was correct for aeon-framework-coupled versions — they referenced memory/ paths and aeon.yml workflows that don't exist for a general user. Reworked to operate on the local installed-skills directory (a pattern the Bankr catalog already uses for skill distribution) rather than aeon-specific paths: - aeon-skill-security-scan — audit installed skills for injection / exfiltration / traversal / prompt-override / obfuscation risks before running them. Delta tracking, baseline suppression, trusted-publisher filter. Designed to feed the Bankr Safety Score concept. - aeon-skill-evals — assertion-based output validation across installed skills. Bootstrap mode generates manifests from real historical runs. Detects regressions before users do. - aeon-skill-repair — diagnose-and-fix loop for a single failing skill. Category-based playbooks (api-change, rate-limit, prompt-bug, etc.), risk classes, verification block attached to every fix. - aeon-autoresearch — genetic-algorithm-style self-improvement. Four parallel variations, weighted rubric, never-downgrade guarantee. Total skill count: 24. The four meta-skills compose with the other 20: - skill-evals surfaces low-quality output - skill-repair fixes deterministic bugs - autoresearch handles quality-lift evolution - skill-security-scan is the input-safety net before any skill runs 🤖 Generated with Claude Code Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
… substance Total cut: 2323 → 1451 lines (-37%). Most skills now in the 50-80 line range, matching the existing Bankr catalog norms (Quotient 75, checkr 107). Common cuts applied: - Removed "Pairs with" sections (10+ skills) — composition is self-evident from the description and frontmatter triggers - Removed multi-paragraph "Why this exists" narrative blocks - Collapsed 5-8-row tables into 2-3 sentences where the table was descriptive rather than reference data - Removed "Required keys: None" lines and long zero-content "Guidelines" lists - Tightened sample outputs from full-document mockups to representative lines - Dropped academic frameworks (4-tier source × 5-confidence in deep-research, 6-axis rubric in narrative-tracker, Stage × Impact matrix in reg-monitor) in favor of the operational core Kept intact: - distribute-tokens (financial precision — idempotency, two-phase, recovery) - monitor-runners (the filters are the value) - vuln-scanner (scanner choices + disclosure routing are the substance) Functional behavior is unchanged; the descriptions are tighter. 🤖 Generated with Claude Code Co-Authored-By: Claude Opus 4.7 (1M context) <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
No description provided.