Skip to content

ccoalm/ccl-skills / skills/skill-extraction-workflow

skill-extraction-workflow

Flagged

复盘 / 沉淀 / 总结经验 / 补进技能 / 技能缺陷 / 流程优化 / would other teammates hit this / will we forget next time / teamwide recurrence / extract skill / post-mortem / 深度 review·审计技能仓库 / 对标外部技能包找 gap(benchmark superpowers·gstack 等) / 核查全局安装点(~/.config/opencode 等)旧快照是否遮蔽本仓技能 / 本仓(ccl-skills 等共享技能仓)OpenCode 项目配置·命令治理 → turn reusable lessons into durable CCL skills.

FileKindSize
SKILL.mdskill md111.0 KB
agents/openai.yamlconfig257 B
references/analysis-parse-fix-test-challenge-replay.mdmarkdown1.5 KB
references/attention-budget-ratchet.mdmarkdown13.7 KB
references/attribution-verification.mdmarkdown4.4 KB
references/bootstrap-slim-c3-obligation-table.mdmarkdown8.0 KB
references/correction-routing-map.mdmarkdown3.7 KB
references/coverage-exhaustion-traps.mdmarkdown19.6 KB
references/description-authoring.mdmarkdown21.1 KB
references/dual-track-review-gate.mdmarkdown110.6 KB
references/eval-routing.mdmarkdown22.2 KB
references/evidence-card-template.mdmarkdown3.2 KB
references/example-domain-preselect.mdmarkdown4.6 KB
references/external-practice-controls.mdmarkdown45.5 KB
references/extraction-lifecycle-handoff.mdmarkdown6.9 KB
references/extraction-quickstart.mdmarkdown20.2 KB
references/firing-point-placement.mdmarkdown21.3 KB
references/harness-patterns-and-eval.mdmarkdown42.6 KB
references/incident-postmortem-extraction.mdmarkdown28.8 KB
references/l0-l1-l2-routing.mdmarkdown10.3 KB
references/online-skill-review.mdmarkdown6.7 KB
references/parallel-stack-references-pattern.mdmarkdown35.9 KB
references/r0-leakage-audit.mdmarkdown19.9 KB
references/recurring-anti-patterns-checklist.mdmarkdown34.4 KB
references/resume-paused-delivery.mdmarkdown3.0 KB
references/review-feedback-mining.mdmarkdown5.7 KB
references/review-finding-standards.mdmarkdown7.3 KB
references/review-rubric.mdmarkdown3.1 KB
references/rule-consolidation.mdmarkdown38.2 KB
references/skill-listing-budget.mdmarkdown1.7 KB
references/source-register.mdmarkdown1.0 MB
references/source-to-skill-extraction.mdmarkdown109.3 KB
references/two-source-extraction-pattern.mdmarkdown19.4 KB
references/uiux-judgment-extraction.mdmarkdown19.1 KB
references/uiux-routing-map.mdmarkdown5.8 KB
references/validation-and-landing.mdmarkdown46.9 KB
scripts/AGENTS.mdmarkdown635 B
scripts/check-ccl-skills.shexecutablescript100.7 KB
scripts/check-contract-anchors.shexecutablescript5.6 KB
scripts/check-evidence-card-leak.shexecutablescript26.1 KB
scripts/check-mr-target-freshness.shexecutablescript9.4 KB
scripts/check-size-budget.shexecutablescript36.0 KB
scripts/check-sync-pointers.shexecutablescript22.6 KB
scripts/check_review_evidence_present.pyexecutablescript4.6 KB
scripts/contract-anchors.tsvother3.8 KB
scripts/entrypoint_form_census.pyexecutablescript7.1 KB
scripts/eval-golden-trace.rbexecutablescript10.0 KB
scripts/eval-health.rbexecutablescript14.6 KB
scripts/eval-routing-bank.rbexecutablescript30.4 KB
scripts/eval-routing.rbexecutablescript9.2 KB
scripts/extraction_review_gate.shexecutablescript1.8 KB
scripts/gate_receipt.pyexecutablescript24.7 KB
scripts/generic-r0-leak-scan.shexecutablescript12.4 KB
scripts/governing-chain-diff.pyexecutablescript12.8 KB
scripts/impact-chain-gate.rbexecutablescript124.4 KB
scripts/obligation-ledger.pyscript107.0 KB
scripts/reference-access-census.shexecutablescript8.3 KB
scripts/register-firing-path-resolution.rbexecutablescript48.4 KB
scripts/shared_git_surface_gate.pyexecutablescript38.0 KB
scripts/skill-behavior-eval.pyexecutablescript39.3 KB
scripts/skill-paired-eval.pyexecutablescript76.9 KB
scripts/source-register-lifecycle.rbscript1.5 KB
scripts/source-register-pending-status.rbscript2.0 KB
scripts/test_ai_coding_implementation_gates.shexecutablescript85.2 KB
scripts/test_antipattern_grep_panel.shexecutablescript3.2 KB
scripts/test_body_compliance_grading.shexecutablescript21.0 KB
scripts/test_check_ccl_impact_chain_refscripts.shexecutablescript145.1 KB
scripts/test_check_ccl_r0_status.shexecutablescript4.2 KB
scripts/test_check_ccl_register_pending_exclusion.shexecutablescript7.4 KB
scripts/test_check_ccl_regressions.shexecutablescript14.3 KB
scripts/test_check_ccl_route_drift.shexecutablescript22.6 KB
scripts/test_check_ccl_size_budget.shexecutablescript59.5 KB
scripts/test_check_ccl_skill_catalog.shexecutablescript23.3 KB
scripts/test_check_ccl_source_register_lifecycle.shexecutablescript6.1 KB
scripts/test_check_contract_anchors.shexecutablescript7.4 KB
scripts/test_check_mr_target_freshness.shexecutablescript12.8 KB
scripts/test_check_review_evidence_present.shexecutablescript4.3 KB
scripts/test_check_sync_pointers.shexecutablescript32.5 KB
scripts/test_ci_checkout_ref_binding.shexecutablescript8.2 KB
scripts/test_controlled_escalation_pins.shexecutablescript7.7 KB
scripts/test_entrypoint_domain_scan_terms.shexecutablescript8.8 KB
scripts/test_entrypoint_form_census.shexecutablescript6.3 KB
scripts/test_eval_routing_bank_grader_diagnostics.shexecutablescript20.6 KB
scripts/test_eval_routing_bank_resolution.shexecutablescript15.3 KB
scripts/test_eval_routing_bank_surface_binding.shexecutablescript6.9 KB
scripts/test_eval_routing_prose_target.shexecutablescript3.7 KB
scripts/test_eval_runtime.pyexecutablescript43.3 KB
scripts/test_extraction_review_gate.shexecutablescript15.2 KB
scripts/test_frozen_case_sanctity.shexecutablescript6.9 KB
scripts/test_frozen_case_sanctity_selfproof.shexecutablescript4.6 KB
scripts/test_gate_receipt.shexecutablescript17.7 KB
scripts/test_generic_r0_leak_scan.shexecutablescript6.4 KB
scripts/test_git_identity_predicate_gate.shscript13.8 KB
scripts/test_governing_chain_diff.shexecutablescript15.1 KB
scripts/test_impact_chain_gate_dateless_host.shexecutablescript6.8 KB
scripts/test_impact_chain_gate_verdict_differential.shexecutablescript20.8 KB
scripts/test_impact_chain_round_attribution.shexecutablescript32.5 KB
scripts/test_impact_chain_self_adjudication.shexecutablescript31.4 KB
scripts/test_impact_chain_source_refuted.shexecutablescript14.7 KB
scripts/test_liveness_predicate_gate.shexecutablescript14.5 KB
scripts/test_locale_independent_gates.shexecutablescript13.0 KB
scripts/test_obligation_ledger.shexecutablescript94.9 KB
scripts/test_obligation_ledger_repo_audit.shexecutablescript2.7 KB
scripts/test_pinned_phrase_mutation_walk.shexecutablescript6.5 KB
scripts/test_reference_access_census.shexecutablescript17.3 KB
scripts/test_register_firing_path_resolution.shexecutablescript41.0 KB
scripts/test_register_firing_path_wiring.shexecutablescript29.9 KB
scripts/test_regression_runner_lanes.shexecutablescript4.5 KB
scripts/test_regression_runner_registration.shexecutablescript1.5 KB
scripts/test_routing_bank_integrity.shexecutablescript15.5 KB
scripts/test_routing_pointer_integrity.shexecutablescript18.9 KB
scripts/test_shared_git_surface_gate.shexecutablescript76.8 KB
scripts/test_skill_paired_eval.pyscript60.0 KB
scripts/test_teardown_guard_pins.shexecutablescript17.0 KB
scripts/test_uiux_delivery_contract.shexecutablescript118.6 KB
scripts/test_uiux_loading_budget.shexecutablescript15.4 KB
scripts/test_validate_skill_credential_cwd.shexecutablescript2.6 KB
scripts/test_validate_skill_cross_refs.shexecutablescript4.8 KB
scripts/test_validate_skill_root_depth.shexecutablescript2.5 KB
scripts/validate-skill.shexecutablescript10.5 KB

scripts/reference-access-census.sh

View on GitHub
#!/usr/bin/env bash
# Reference-access census: which of a skill package's files were actually
# touched by real agent sessions on THIS host, over a lookback window.
#
# Why this exists: a rule set must not grow monotonically, but "retire what is
# not pulling its weight" needs a signal that is not the author's opinion. The
# closest observable proxy on a local host is the per-file access record in the
# agent's own session transcripts (Claude Code `~/.claude/projects/**/*.jsonl`,
# Codex `~/.codex/sessions/**/*.jsonl`): a reference that no session mentioned
# in N days is a relocation/retirement candidate; one mentioned in most sessions
# that loaded the skill is a candidate for promotion into the entrypoint. This
# is the local instantiation of the usage counters that context-evolution
# methods keep per bullet (helpful/harmful counts) and of the "ignored content
# is unnecessary or poorly signaled" observation in the official skill-authoring
# guidance — advisory only, never a gate (Goodhart: a count that becomes a
# target gets gamed by mentioning files).
#
# Privacy contract: the transcripts are private per-host data. This script
# prints ONLY repo-relative skill file paths, per-file session counts, and
# last-touched dates. It never prints transcript text, prompts, absolute paths
# outside the skill tree, or session ids. Its output is safe to paste into a
# private charter; it is still not shared-tree content by itself.
#
# Counting unit: a SESSION (one transcript file) counts once per skill file it
# mentions, whatever the tool (Read, sed, grep, Skill load). "Mentioned" is a
# superset of "read to depth" — treat a count as an upper bound on real use.
#
# Usage:
#   reference-access-census.sh [--skill <name>] [--days <n>] [--repo-root <dir>]
#                              [--logs <dir>[,<dir>...]]
# Defaults: skill=skill-extraction-workflow, days=60, repo-root=cwd-derived,
#           logs=$HOME/.claude/projects,$HOME/.codex/sessions
# Exit 0 on a completed census or an honest unevaluated result (no transcripts);
# exit 2 on usage errors and on input errors (unreadable/vanished/unexecutable
# inputs) — counts are withheld rather than printed as zeros.
set -euo pipefail

skill="skill-extraction-workflow"
days=60
repo_root=""
logs="${HOME:+${HOME}/.claude/projects,${HOME}/.codex/sessions}"
logs_explicit=0
while [ $# -gt 0 ]; do
  case "$1" in
    --skill|--days|--repo-root|--logs)
      if [ $# -lt 2 ]; then echo "reference_access_census_usage_error: $1 needs a value" >&2; exit 2; fi ;;
  esac
  case "$1" in
    --skill) skill="$2"; shift 2 ;;
    --days) days="$2"; shift 2 ;;
    --repo-root) repo_root="$2"; shift 2 ;;
    --logs) logs="$2"; logs_explicit=1; shift 2 ;;
    -h|--help) sed -n '2,32p' "$0"; exit 0 ;;
    *) echo "reference_access_census_usage_error: unknown argument (see --help)" >&2; exit 2 ;;
  esac
done
if [ -z "$repo_root" ]; then
  repo_root="$(git rev-parse --show-toplevel 2>/dev/null || pwd)"
fi
skill_dir="$repo_root/skills/$skill"
if [ ! -d "$skill_dir" ]; then
  echo "reference_access_census_usage_error: --skill names no tracked skill package under the repo root" >&2; exit 2
fi
case "$days" in ''|*[!0-9]*) echo "reference_access_census_usage_error: --days must be an integer" >&2; exit 2 ;; esac

# Inventory: every tracked markdown file in the package (entrypoint + references).
files=()
while IFS= read -r f; do files+=("$f"); done < <(cd "$repo_root" && git ls-files "skills/$skill/SKILL.md" "skills/$skill/references/*.md" 2>/dev/null | sort)
if [ "${#files[@]}" -eq 0 ]; then
  echo "reference_access_census_usage_error: --skill names a package with no tracked SKILL.md or references" >&2; exit 2
fi

# Candidate transcripts: any .jsonl under the log roots modified within the window.
tmp="$(mktemp -d)"; trap 'rm -rf "$tmp"' EXIT
roots=(); if [ -n "$logs" ]; then IFS=',' read -r -a roots <<<"$logs"; fi
: > "$tmp/candidates0"; : > "$tmp/errors"
# Input errors are never zeros: every scan phase collects stderr, and a non-empty
# error log withholds the table (exit 2) instead of printing counts that would
# read as evidence. grep's no-match status writes nothing to stderr, so it never
# trips this; unreadable, vanished, or unexecutable inputs do. The error log is
# summarized as a count only — file names would be private paths.
withhold_if_errors() {
  if [ -s "$tmp/errors" ]; then
    n="$(wc -l < "$tmp/errors" | tr -d ' ')"
    echo "reference_access_census_unevaluated: $n input error(s) while scanning transcripts (unreadable, vanished, or unexecutable inputs); counts withheld"
    exit 2
  fi
}
# A default root that is absent is normal (the host may run only one agent); an
# explicitly supplied root that is absent is an input error — the caller asked
# for a scan that cannot happen, and skipping it would read as a real count.
for r in "${roots[@]}"; do
  [ -n "$r" ] || continue
  if [ ! -d "$r" ]; then
    if [ "$logs_explicit" -eq 1 ]; then echo "missing explicit log root" >> "$tmp/errors"; fi
    continue
  fi
  find "$r" -type f -name '*.jsonl' -mtime "-$days" -print0 >> "$tmp/candidates0" 2>> "$tmp/errors" || echo "transcript listing failed under one log root" >> "$tmp/errors"
done
withhold_if_errors
# The inventory is NUL-delimited (a pathname may contain a newline) and
# deduplicated by pathname: overlapping roots (a root and its own subtree, or
# the same root twice) list a transcript more than once, and a session must
# count once. Two names for one file (symlinked roots) are not canonicalized.
sort -zu "$tmp/candidates0" -o "$tmp/candidates0"
total_sessions="$(tr -cd '\0' < "$tmp/candidates0" | wc -c | tr -d ' ')"
if [ "$total_sessions" -eq 0 ]; then
  echo "reference_access_census_unevaluated: no transcripts within ${days}d under ${#roots[@]} log root(s); counts withheld"
  exit 0
fi

# Sessions that mention the package at all (fixed-string prefilter keeps this fast).
# xargs batches keep this under ARG_MAX; grep exit 1 (no match in a batch) is not an error.
# grep exit 1 (no match in a batch) is success; any other non-zero status —
# with or without a stderr line — is an input error, so each batch runs under a
# small wrapper that normalizes 1 to 0 and lets xargs report anything else.
# `|| rc=$?` keeps the batch alive when the caller exported SHELLOPTS=errexit:
# a bare `grep; rc=$?` would exit the child at status 1 before normalizing it.
grep_batch() { local rc=0; grep -lF --null "$@" || rc=$?; [ "$rc" -le 1 ] && return 0; return "$rc"; }
export -f grep_batch 2>/dev/null || true
if ! xargs -0 -n 200 bash -c 'grep_batch "$@"' _ "skills/$skill/" < "$tmp/candidates0" > "$tmp/touching0" 2>> "$tmp/errors"; then
  echo "transcript scan failed in at least one batch" >> "$tmp/errors"
fi
withhold_if_errors
touching="$(tr -cd '\0' < "$tmp/touching0" | wc -c | tr -d ' ')"

# Per-file session count + last-touched date (mtime of the newest touching transcript).
: > "$tmp/rows"
for f in "${files[@]}"; do
  rel="${f#skills/$skill/}"
  if [ "$touching" -eq 0 ]; then n=0; last="-"; else
    if ! xargs -0 -n 200 bash -c 'grep_batch "$@"' _ "$f" < "$tmp/touching0" > "$tmp/hits0" 2>> "$tmp/errors"; then
      echo "transcript scan failed in at least one batch" >> "$tmp/errors"
    fi
    withhold_if_errors
    n="$(tr -cd '\0' < "$tmp/hits0" | wc -c | tr -d ' ')"
    if [ "$n" -gt 0 ]; then
      # BSD stat first, GNU stat as the fallback; only the fallback's failure is an
      # error. Every substitution is guarded so a failing pipeline cannot abort the
      # script under set -e before withhold_if_errors runs.
      last="$(xargs -0 stat -f '%Sm' -t '%Y-%m-%d' < "$tmp/hits0" 2>/dev/null | sort | tail -1)" || last=""
      if [ -z "$last" ]; then
        last="$(xargs -0 stat -c '%y' < "$tmp/hits0" 2>> "$tmp/errors" | cut -c1-10 | sort | tail -1)" || last=""
      fi
      if [ -z "$last" ]; then
        [ -s "$tmp/errors" ] || echo "stat produced no timestamp for a touching transcript" >> "$tmp/errors"
        withhold_if_errors
      fi
    else last="-"; fi
  fi
  if [ "$touching" -gt 0 ]; then share="$(( n * 100 / touching ))%"; else share="-"; fi
  printf '%s | %s | %s | %s\n' "$rel" "$n" "$last" "$share" >> "$tmp/rows"
done
printf '%s\n' "reference_access_census: skill=$skill window=${days}d transcripts=$total_sessions sessions_touching_package=$touching"
printf '%s\n' "file | sessions | last_touched | share_of_touching"
sort -t'|' -k2,2nr "$tmp/rows"
echo "reference_access_census_ok"
reverse-engineer-anythingmorluto/reaReverse engineer native, managed, Electron/JavaScript, packaged, firmware, and browser targets with REA, and build agent environments (AgentEnv bundles) of web applications from a live instance or a capture, and of Linux desktop apps with a task from a request. Use shipped-artifact or requested runtime evidence to explain features, compare versions, decompile code, or guide a reconstruction. Skip REA for ordinary source-repository architecture analysis.Stars: 471kUpdated Claude CodeCodexVerifiedInstall
diagnosing-superpowersobra/superpowersUse when a superpowers session went wrong and your human partner wants to know why — repeated work, ignored plans, stumbles, poor results, a skill that didn't fire, "it took too long", "why is it so expensive", "what is it doing" — or wants to build a bug report for the superpowers maintainers, for the current session or a past one identified by id or path, on any harness.Stars: 322kUpdated Claude CodeCodexVerifiedInstall
dispatching-parallel-agentsobra/superpowersUse when facing 2+ independent tasks that can be worked on without shared state or sequential dependenciesStars: 322kUpdated Claude CodeCodexVerifiedInstall
executing-plansobra/superpowersUse when executing an implementation plan in the current session as the implementer yourself — your human partner chose inline execution, or no subagent tool is availableStars: 322kUpdated Claude CodeCodexVerifiedInstall